Executive summary
vosen/ZLUDA falls short of our standard for production, held back most by its Code Health lens, the quality of the code itself. It scores 55 out of 100 and is rated Adequate, and 46 of its 48 critical findings are security problems in the build pipeline and in the packages it uses.
Code Health, at 46, is the weakest lens and counts most in the score. The best first step for the work it takes is recording design decisions, worth about 5 points. Turning on missing-value checks would lift Code Health itself, worth about 8 points. Raising Code Health to 70 would lift the score to about 61.
Forty-one pipeline tools are referenced by a changeable name, so the build pipeline can be tampered with. Two pipeline steps run downloads nobody checks, and two secrets can leak from the build pipeline. Pinning each tool to an exact version, checking each download against a checksum and limiting each secret to its own step would close these risks. Updating the vulnerable packages would fix the package with a known security hole. The part of the code nobody has worked on for a long time needs someone to learn and document it before anyone changes it.
The product code is 86,872 lines, about €89,000 to rebuild.
If Code Health rose from 46 to 70, the lowest score rated Strong, the score would rise from 55 to about 61.
New since the last survey: 28 findings
- Cognitive Complexity · zluda::impl::surf::convert_resource_desc (cognitive 19) zluda/src/impl/surf.rs
- Cognitive Complexity · ptx_parser::multi_variable (cognitive 17) ptx_parser/src/lib.rs
- God Classes · FileTooLong: src/lib.rs zluda_windows/src/lib.rs
- Code Duplication · Edited copy of a member (16 corresponding lines) format/src/lib.rs
- Code Duplication · Members sharing a duplicated core (5 members, 50+ identical tokens) zluda_bindgen/src/main.rs
- Code Duplication · Duplicated block (22 lines × 2) zluda/src/impl/surf.rs
- Code Duplication · Duplicated block (12–13 lines × 2) zluda/src/impl/graphics.rs
- Code Duplication · Duplicated block (6 lines × 2) zluda/src/impl/driver.rs
- Dependency Vulnerabilities · REDACTED
- Dependency Vulnerabilities · REDACTED
- Dependency Vulnerabilities · REDACTED
- Dependency Vulnerabilities · REDACTED
- Dependency Vulnerabilities · REDACTED
- Dependency Vulnerabilities · REDACTED
- Dependency Vulnerabilities · REDACTED
- Dependency Vulnerabilities · REDACTED
- Dependency Vulnerabilities · REDACTED
- Dependency Vulnerabilities · REDACTED
- Dependency Vulnerabilities · REDACTED
- Dependency Vulnerabilities · REDACTED
- Dependency Vulnerabilities · REDACTED
- Dependency Vulnerabilities · REDACTED
- Dependency Vulnerabilities · REDACTED
- Change Coupling · Change coupling: function.rs ↔ memory.rs zluda/src/impl/function.rs
- Supply-chain Provenance & Signing · REDACTED
- Supply-chain Provenance & Signing · REDACTED
- Supply-chain Provenance & Signing · REDACTED
- Supply-chain Provenance & Signing · REDACTED
Fixed since the last survey: 3 findings
- Code Duplication · Edited copy of a member (16 corresponding lines) format/src/lib.rs
- Code Duplication · Members sharing a duplicated core (5 members, 50+ identical tokens) zluda_bindgen/src/main.rs
- Supply-chain Provenance & Signing REDACTED
Rebuild cost: about €89,000 · Estimate
| Cost to rebuild | €30,000 to €150,000, for 494 to 1,569 hours of work, about 0.6 person-years for 1 to 2 engineers |
|---|---|
| Difficulty of the problem | Standard, judged from the kind of software and how it is built. Code that solves a harder problem, such as payments or medical bookings, costs more per line to write. |
| Quality factor | ×0.8 at a score of 55. Better code takes more work to write, so the factor is ×1.0 at a score of 70, lower below it and higher above it. |
| Size | Medium: 86,872 lines of product code |
| What it is not | Not a quotation: nobody has offered to build it for that. Not a valuation: it says nothing about what the business is worth. Not a certified figure: no professional has signed it as one. |
Top priorities
Diagnosis: what the measurements show together
Architecture — module dependency graph
Architecture: which parts of the code depend on which
| uses → | 1 kernel_metadata | 2 llvm_zluda.utils | 3 zluda_common | 4 cuda_check.win | 5 format | 6 hip_runtime_sys | 7 ptx_parser.ast | 8 cuda_types.cublaslt | 9 cuda_types.cudnn | 10 cuda_types.cufft | 11 cuda_types.nvml | 12 dark_api | 13 ptx_parser_macros_impl | 14 cuda_types.cudnn9 | 15 cuda_types.cusparse | 16 dark_api.cuda | 17 hipblaslt_sys | 18 ptx_parser_macros_impl.parser | 19 cuda_types.cublas | 20 miopen_sys | 21 ptx_parser | 22 zluda_trace | 23 cuda_types.cudnn8 | 24 ptx.pass.instruction_mode_to_global_mode | 25 zluda_dnn.impl | 26 zluda_trace.log | 27 ptx.pass.llvm.emit | 28 zluda_server_common | 29 cuda_types.cuda | 30 ptx.pass | 31 ptx.pass.insert_explicit_load_store | 32 zluda.impl.context | 33 zluda.impl.function | 34 zluda.impl.library | 35 compiler.hip | 36 zluda.impl.graphics | 37 zluda_blas.impl | 38 zluda.impl.driver | 39 zluda.impl.module | 40 zluda_blaslt.impl |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 kernel_metadata | ||||||||||||||||||||||||||||||||||||||||
| 2 llvm_zluda.utils | ||||||||||||||||||||||||||||||||||||||||
| 3 zluda_common | ||||||||||||||||||||||||||||||||||||||||
| 4 cuda_check.win | 1 | |||||||||||||||||||||||||||||||||||||||
| 5 format | 3 | |||||||||||||||||||||||||||||||||||||||
| 6 hip_runtime_sys | 4 | 2 | ||||||||||||||||||||||||||||||||||||||
| 7 ptx_parser.ast | 60 | 2 | ||||||||||||||||||||||||||||||||||||||
| 8 cuda_types.cublaslt | 1 | 20 | ||||||||||||||||||||||||||||||||||||||
| 9 cuda_types.cudnn | 56 | |||||||||||||||||||||||||||||||||||||||
| 10 cuda_types.cufft | 1 | 13 | ||||||||||||||||||||||||||||||||||||||
| 11 cuda_types.nvml | 1 | 203 | ||||||||||||||||||||||||||||||||||||||
| 12 dark_api | 1 | |||||||||||||||||||||||||||||||||||||||
| 13 ptx_parser_macros_impl | 2 | |||||||||||||||||||||||||||||||||||||||
| 14 cuda_types.cudnn9 | 7 | 9 | 1 | |||||||||||||||||||||||||||||||||||||
| 15 cuda_types.cusparse | 1 | 26 | 4 | |||||||||||||||||||||||||||||||||||||
| 16 dark_api.cuda | 1 | |||||||||||||||||||||||||||||||||||||||
| 17 hipblaslt_sys | 7 | 4 | 2 | |||||||||||||||||||||||||||||||||||||
| 18 ptx_parser_macros_impl.parser | 1 | 1 | ||||||||||||||||||||||||||||||||||||||
| 19 cuda_types.cublas | 2 | 11 | 1 | |||||||||||||||||||||||||||||||||||||
| 20 miopen_sys | 7 | 5 | 1 | |||||||||||||||||||||||||||||||||||||
| 21 ptx_parser | 1 | 1 | 4 | 1 | ||||||||||||||||||||||||||||||||||||
| 22 zluda_trace | 1 | 2 | 6 | 4 | ||||||||||||||||||||||||||||||||||||
| 23 cuda_types.cudnn8 | 1 | 8 | 9 | 1 | ||||||||||||||||||||||||||||||||||||
| 24 ptx.pass.instruction_mode_to_global_mode | 3 | 1 | 1 | 19 | ||||||||||||||||||||||||||||||||||||
| 25 zluda_dnn.impl | 1 | 2 | 2 | 5 | 1 | 25 | ||||||||||||||||||||||||||||||||||
| 26 zluda_trace.log | 2 | |||||||||||||||||||||||||||||||||||||||
| 27 ptx.pass.llvm.emit | 4 | 44 | 1 | 9 | 1 | 19 | ||||||||||||||||||||||||||||||||||
| 28 zluda_server_common | 1 | 1 | ||||||||||||||||||||||||||||||||||||||
| 29 cuda_types.cuda | 3 | 238 | 1 | 5 | ||||||||||||||||||||||||||||||||||||
| 30 ptx.pass | 2 | 3 | 8 | 8 | 1 | 1 | ||||||||||||||||||||||||||||||||||
| 31 ptx.pass.insert_explicit_load_store | 6 | 1 | 1 | 6 | ||||||||||||||||||||||||||||||||||||
| 32 zluda.impl.context | 1 | 2 | 2 | 2 | ||||||||||||||||||||||||||||||||||||
| 33 zluda.impl.function | 1 | 2 | 1 | 1 | 1 | |||||||||||||||||||||||||||||||||||
| 34 zluda.impl.library | 1 | 2 | 1 | |||||||||||||||||||||||||||||||||||||
| 35 compiler.hip | 1 | |||||||||||||||||||||||||||||||||||||||
| 36 zluda.impl.graphics | 1 | 1 | 2 | 1 | ||||||||||||||||||||||||||||||||||||
| 37 zluda_blas.impl | 2 | 1 | 2 | 1 | ||||||||||||||||||||||||||||||||||||
| 38 zluda.impl.driver | 1 | 2 | 5 | 1 | 1 | |||||||||||||||||||||||||||||||||||
| 39 zluda.impl.module | 2 | 1 | 2 | 2 | 1 | 2 | 1 | 1 | ||||||||||||||||||||||||||||||||
| 40 zluda_blaslt.impl | 3 | 1 | 4 | 8 | 1 |
4→34 cuda_check.win uses zluda.impl.library part of a loop✕
- Connections
- 1 pair of types, where a type in cuda_check.win uses a type in zluda.impl.library.
- The types involved
cuda_check.win.win$module→zluda.impl.library.Library
- Direction
- cuda_check.win uses zluda.impl.library, so a change to zluda.impl.library can break cuda_check.win, while a change to cuda_check.win can't break zluda.impl.library.
- Position
- Above the diagonal, so it is part of a loop. Neither module can be changed, tested or released on its own until one of these dependencies is removed.
5→29 format uses cuda_types.cuda part of a loop✕
- Connections
- 3 pairs of types, where a type in format uses a type in cuda_types.cuda.
- The types involved
format.format$module→cuda_types.cuda.CUDA_EXTERNAL_MEMORY_HANDLE_DESC_st__bindgen_ty_1__bindgen_ty_1format.format$module→cuda_types.cuda.CUstreamformat.format$module→cuda_types.cuda.CUstreamBatchMemOpParams_union_CUstreamMemOpWaitValueParams_st
- Direction
- format uses cuda_types.cuda, so a change to cuda_types.cuda can break format, while a change to format can't break cuda_types.cuda.
- Position
- Above the diagonal, so it is part of a loop. Neither module can be changed, tested or released on its own until one of these dependencies is removed.
6→3 hip_runtime_sys uses zluda_common✕
- Connections
- 4 pairs of types, where a type in hip_runtime_sys uses a type in zluda_common.
- The types involved
hip_runtime_sys.HIP_MEMCPY3D→zluda_common.FromCudahip_runtime_sys.hipJitOption→zluda_common.FromCudahip_runtime_sys.hipLimit_t→zluda_common.FromCudahip_runtime_sys.hip_Memcpy2D→zluda_common.FromCuda
- Direction
- hip_runtime_sys uses zluda_common, so a change to zluda_common can break hip_runtime_sys, while a change to hip_runtime_sys can't break zluda_common.
- Position
- Below the diagonal: hip_runtime_sys uses a module listed above it, which is the healthy direction.
6→35 hip_runtime_sys uses compiler.hip part of a loop✕
- Connections
- 2 pairs of types, where a type in hip_runtime_sys uses a type in compiler.hip.
- The types involved
hip_runtime_sys.hipDeviceArch_t→compiler.hip.__BindgenBitfieldUnithip_runtime_sys.hipDeviceProp_tR0600→compiler.hip.hipDeviceArch_t
- Direction
- hip_runtime_sys uses compiler.hip, so a change to compiler.hip can break hip_runtime_sys, while a change to hip_runtime_sys can't break compiler.hip.
- Position
- Above the diagonal, so it is part of a loop. Neither module can be changed, tested or released on its own until one of these dependencies is removed.
7→21 ptx_parser.ast uses ptx_parser part of a loop✕
- Connections
- 60 pairs of types, where a type in ptx_parser.ast uses a type in ptx_parser. Showing 25 of them; the rest are in namespace-graph.json in this report's bundle.
- The types involved
ptx_parser.ast.ArithDetails→ptx_parser.ScalarTypeptx_parser.ast.ArithFloat→ptx_parser.ScalarTypeptx_parser.ast.ArithInteger→ptx_parser.ScalarTypeptx_parser.ast.AtomCasDetails→ptx_parser.AtomSemanticsptx_parser.ast.AtomCasDetails→ptx_parser.MemScopeptx_parser.ast.AtomCasDetails→ptx_parser.ScalarTypeptx_parser.ast.AtomCasDetails→ptx_parser.StateSpaceptx_parser.ast.AtomDetails→ptx_parser.AtomSemanticsptx_parser.ast.AtomDetails→ptx_parser.MemScopeptx_parser.ast.AtomDetails→ptx_parser.StateSpaceptx_parser.ast.BarRedData→ptx_parser.Reductionptx_parser.ast.CarryDetails→ptx_parser.ScalarTypeptx_parser.ast.CpAsyncDetails→ptx_parser.StateSpaceptx_parser.ast.CreatePolicyFractionalDetails→ptx_parser.EvictionPriorityptx_parser.ast.CvtDesc→ptx_parser.ScalarTypeptx_parser.ast.CvtDetails→ptx_parser.PtxErrorptx_parser.ast.CvtDetails→ptx_parser.ScalarTypeptx_parser.ast.CvtIntToIntDesc→ptx_parser.ScalarTypeptx_parser.ast.CvtaDetails→ptx_parser.StateSpaceptx_parser.ast.DivDetails→ptx_parser.ScalarTypeptx_parser.ast.DivFloatDetails→ptx_parser.ScalarTypeptx_parser.ast.Dp2aData→ptx_parser.ScalarTypeptx_parser.ast.Dp4aDetails→ptx_parser.ScalarTypeptx_parser.ast.LdDetails→ptx_parser.StateSpaceptx_parser.ast.LdMatrixDetails→ptx_parser.MatrixNumber
- Direction
- ptx_parser.ast uses ptx_parser, so a change to ptx_parser can break ptx_parser.ast, while a change to ptx_parser.ast can't break ptx_parser.
- Position
- Above the diagonal, so it is part of a loop. Neither module can be changed, tested or released on its own until one of these dependencies is removed.
7→24 ptx_parser.ast uses ptx.pass.instruction_mode_to_global_mode part of a loop✕
- Connections
- 2 pairs of types, where a type in ptx_parser.ast uses a type in ptx.pass.instruction_mode_to_global_mode.
- The types involved
ptx_parser.ast.ArithFloat→ptx.pass.instruction_mode_to_global_mode.RoundingModeptx_parser.ast.CvtDesc→ptx.pass.instruction_mode_to_global_mode.RoundingMode
- Direction
- ptx_parser.ast uses ptx.pass.instruction_mode_to_global_mode, so a change to ptx.pass.instruction_mode_to_global_mode can break ptx_parser.ast, while a change to ptx_parser.ast can't break ptx.pass.instruction_mode_to_global_mode.
- Position
- Above the diagonal, so it is part of a loop. Neither module can be changed, tested or released on its own until one of these dependencies is removed.
8→3 cuda_types.cublaslt uses zluda_common✕
- Connections
- 1 pair of types, where a type in cuda_types.cublaslt uses a type in zluda_common.
- The types involved
cuda_types.cublaslt.cublasLtMatmulHeuristicResult_t→zluda_common.FromCuda
- Direction
- cuda_types.cublaslt uses zluda_common, so a change to zluda_common can break cuda_types.cublaslt, while a change to cuda_types.cublaslt can't break zluda_common.
- Position
- Below the diagonal: cuda_types.cublaslt uses a module listed above it, which is the healthy direction.
8→5 cuda_types.cublaslt uses format✕
- Connections
- 20 pairs of types, where a type in cuda_types.cublaslt uses a type in format.
- The types involved
cuda_types.cublaslt.cublasLtBatchMode_t→format.CudaDisplaycuda_types.cublaslt.cublasLtClusterShape_t→format.CudaDisplaycuda_types.cublaslt.cublasLtEpilogue_t→format.CudaDisplaycuda_types.cublaslt.cublasLtMatmulAlgoCapAttributes_t→format.CudaDisplaycuda_types.cublaslt.cublasLtMatmulAlgoConfigAttributes_t→format.CudaDisplaycuda_types.cublaslt.cublasLtMatmulAlgo_t→format.CudaDisplaycuda_types.cublaslt.cublasLtMatmulDescAttributes_t→format.CudaDisplaycuda_types.cublaslt.cublasLtMatmulHeuristicResult_t→format.CudaDisplaycuda_types.cublaslt.cublasLtMatmulInnerShape_t→format.CudaDisplaycuda_types.cublaslt.cublasLtMatmulMatrixScale_t→format.CudaDisplaycuda_types.cublaslt.cublasLtMatmulPreferenceAttributes_t→format.CudaDisplaycuda_types.cublaslt.cublasLtMatmulSearch_t→format.CudaDisplaycuda_types.cublaslt.cublasLtMatmulStages_t→format.CudaDisplaycuda_types.cublaslt.cublasLtMatmulTile_t→format.CudaDisplaycuda_types.cublaslt.cublasLtMatrixLayoutAttribute_t→format.CudaDisplaycuda_types.cublaslt.cublasLtMatrixTransformDescAttributes_t→format.CudaDisplaycuda_types.cublaslt.cublasLtOrder_t→format.CudaDisplaycuda_types.cublaslt.cublasLtPointerModeMask_t→format.CudaDisplaycuda_types.cublaslt.cublasLtPointerMode_t→format.CudaDisplaycuda_types.cublaslt.cublasLtReductionScheme_t→format.CudaDisplay
- Direction
- cuda_types.cublaslt uses format, so a change to format can break cuda_types.cublaslt, while a change to cuda_types.cublaslt can't break format.
- Position
- Below the diagonal: cuda_types.cublaslt uses a module listed above it, which is the healthy direction.
9→5 cuda_types.cudnn uses format✕
- Connections
- 56 pairs of types, where a type in cuda_types.cudnn uses a type in format. Showing 25 of them; the rest are in namespace-graph.json in this report's bundle.
- The types involved
cuda_types.cudnn.cudnnActivationMode_t→format.CudaDisplaycuda_types.cudnn.cudnnBackendAttributeType_t→format.CudaDisplaycuda_types.cudnn.cudnnBackendHeurMode_t→format.CudaDisplaycuda_types.cudnn.cudnnBackendLayoutType_t→format.CudaDisplaycuda_types.cudnn.cudnnBackendNormFwdPhase_t→format.CudaDisplaycuda_types.cudnn.cudnnBatchNormMode_t→format.CudaDisplaycuda_types.cudnn.cudnnBatchNormOps_t→format.CudaDisplaycuda_types.cudnn.cudnnBnFinalizeStatsMode_t→format.CudaDisplaycuda_types.cudnn.cudnnCTCLossAlgo_t→format.CudaDisplaycuda_types.cudnn.cudnnConvolutionBwdDataAlgo_t→format.CudaDisplaycuda_types.cudnn.cudnnConvolutionBwdFilterAlgo_t→format.CudaDisplaycuda_types.cudnn.cudnnConvolutionFwdAlgo_t→format.CudaDisplaycuda_types.cudnn.cudnnConvolutionMode_t→format.CudaDisplaycuda_types.cudnn.cudnnDeterminism_t→format.CudaDisplaycuda_types.cudnn.cudnnDirectionMode_t→format.CudaDisplaycuda_types.cudnn.cudnnDivNormMode_t→format.CudaDisplaycuda_types.cudnn.cudnnErrQueryMode_t→format.CudaDisplaycuda_types.cudnn.cudnnFoldingDirection_t→format.CudaDisplaycuda_types.cudnn.cudnnForwardMode_t→format.CudaDisplaycuda_types.cudnn.cudnnFractionStruct→format.CudaDisplaycuda_types.cudnn.cudnnFusedOpsConstParamLabel_t→format.CudaDisplaycuda_types.cudnn.cudnnFusedOpsPointerPlaceHolder_t→format.CudaDisplaycuda_types.cudnn.cudnnFusedOpsVariantParamLabel_t→format.CudaDisplaycuda_types.cudnn.cudnnFusedOps_t→format.CudaDisplaycuda_types.cudnn.cudnnGenStatsMode_t→format.CudaDisplay
- Direction
- cuda_types.cudnn uses format, so a change to format can break cuda_types.cudnn, while a change to cuda_types.cudnn can't break format.
- Position
- Below the diagonal: cuda_types.cudnn uses a module listed above it, which is the healthy direction.
10→3 cuda_types.cufft uses zluda_common✕
- Connections
- 1 pair of types, where a type in cuda_types.cufft uses a type in zluda_common.
- The types involved
cuda_types.cufft.cufftError_t→zluda_common.CudaErrorType
- Direction
- cuda_types.cufft uses zluda_common, so a change to zluda_common can break cuda_types.cufft, while a change to cuda_types.cufft can't break zluda_common.
- Position
- Below the diagonal: cuda_types.cufft uses a module listed above it, which is the healthy direction.
10→5 cuda_types.cufft uses format✕
- Connections
- 13 pairs of types, where a type in cuda_types.cufft uses a type in format.
- The types involved
cuda_types.cufft.cudaLibXtDesc_t→format.CudaDisplaycuda_types.cufft.cudaXtDesc_t→format.CudaDisplaycuda_types.cufft.cufftCompatibility_t→format.CudaDisplaycuda_types.cufft.cufftHandle→format.CudaDisplaycuda_types.cufft.cufftProperty_t→format.CudaDisplaycuda_types.cufft.cufftType_t→format.CudaDisplaycuda_types.cufft.cufftXt1dFactors_t→format.CudaDisplaycuda_types.cufft.cufftXtCallbackType_t→format.CudaDisplaycuda_types.cufft.cufftXtCopyType_t→format.CudaDisplaycuda_types.cufft.cufftXtQueryType_t→format.CudaDisplaycuda_types.cufft.cufftXtSubFormat_t→format.CudaDisplaycuda_types.cufft.cufftXtWorkAreaPolicy_t→format.CudaDisplaycuda_types.cufft.libFormat_t→format.CudaDisplay
- Direction
- cuda_types.cufft uses format, so a change to format can break cuda_types.cufft, while a change to cuda_types.cufft can't break format.
- Position
- Below the diagonal: cuda_types.cufft uses a module listed above it, which is the healthy direction.
11→3 cuda_types.nvml uses zluda_common✕
- Connections
- 1 pair of types, where a type in cuda_types.nvml uses a type in zluda_common.
- The types involved
cuda_types.nvml.nvmlError_t→zluda_common.CudaErrorType
- Direction
- cuda_types.nvml uses zluda_common, so a change to zluda_common can break cuda_types.nvml, while a change to cuda_types.nvml can't break zluda_common.
- Position
- Below the diagonal: cuda_types.nvml uses a module listed above it, which is the healthy direction.
11→5 cuda_types.nvml uses format✕
- Connections
- 203 pairs of types, where a type in cuda_types.nvml uses a type in format. Showing 25 of them; the rest are in namespace-graph.json in this report's bundle.
- The types involved
cuda_types.nvml.nvmlAccountingStats_st→format.CudaDisplaycuda_types.nvml.nvmlActiveVgpuInstanceInfo_v1_t→format.CudaDisplaycuda_types.nvml.nvmlBAR1Memory_st→format.CudaDisplaycuda_types.nvml.nvmlBrandType_enum→format.CudaDisplaycuda_types.nvml.nvmlBridgeChipHierarchy_st→format.CudaDisplaycuda_types.nvml.nvmlBridgeChipInfo_st→format.CudaDisplaycuda_types.nvml.nvmlBridgeChipType_enum→format.CudaDisplaycuda_types.nvml.nvmlC2cModeInfo_v1_t→format.CudaDisplaycuda_types.nvml.nvmlClkMonFaultInfo_struct→format.CudaDisplaycuda_types.nvml.nvmlClkMonStatus_status→format.CudaDisplaycuda_types.nvml.nvmlClockId_enum→format.CudaDisplaycuda_types.nvml.nvmlClockLimitId_enum→format.CudaDisplaycuda_types.nvml.nvmlClockOffset_v1_t→format.CudaDisplaycuda_types.nvml.nvmlClockType_enum→format.CudaDisplaycuda_types.nvml.nvmlComputeInstanceInfo_st→format.CudaDisplaycuda_types.nvml.nvmlComputeInstancePlacement_st→format.CudaDisplaycuda_types.nvml.nvmlComputeInstanceProfileInfo_st→format.CudaDisplaycuda_types.nvml.nvmlComputeInstanceProfileInfo_v2_st→format.CudaDisplaycuda_types.nvml.nvmlComputeInstanceProfileInfo_v3_st→format.CudaDisplaycuda_types.nvml.nvmlComputeMode_enum→format.CudaDisplaycuda_types.nvml.nvmlConfComputeGetKeyRotationThresholdInfo_st→format.CudaDisplaycuda_types.nvml.nvmlConfComputeGpuAttestationReport_st→format.CudaDisplaycuda_types.nvml.nvmlConfComputeGpuCertificate_st→format.CudaDisplaycuda_types.nvml.nvmlConfComputeMemSizeInfo_st→format.CudaDisplaycuda_types.nvml.nvmlConfComputeSetKeyRotationThresholdInfo_st→format.CudaDisplay
- Direction
- cuda_types.nvml uses format, so a change to format can break cuda_types.nvml, while a change to cuda_types.nvml can't break format.
- Position
- Below the diagonal: cuda_types.nvml uses a module listed above it, which is the healthy direction.
12→5 dark_api uses format✕
- Connections
- 1 pair of types, where a type in dark_api uses a type in format.
- The types involved
dark_api.FunctionArgInfo→format.CudaDisplay
- Direction
- dark_api uses format, so a change to format can break dark_api, while a change to dark_api can't break format.
- Position
- Below the diagonal: dark_api uses a module listed above it, which is the healthy direction.
13→7 ptx_parser_macros_impl uses ptx_parser.ast✕
- Connections
- 2 pairs of types, where a type in ptx_parser_macros_impl uses a type in ptx_parser.ast.
- The types involved
ptx_parser_macros_impl.ArgumentField→ptx_parser.ast.Typeptx_parser_macros_impl.InstructionVariant→ptx_parser.ast.Type
- Direction
- ptx_parser_macros_impl uses ptx_parser.ast, so a change to ptx_parser.ast can break ptx_parser_macros_impl, while a change to ptx_parser_macros_impl can't break ptx_parser.ast.
- Position
- Below the diagonal: ptx_parser_macros_impl uses a module listed above it, which is the healthy direction.
14→5 cuda_types.cudnn9 uses format✕
- Connections
- 7 pairs of types, where a type in cuda_types.cudnn9 uses a type in format.
- The types involved
cuda_types.cudnn9.cudnnBackendAttributeName_t→format.CudaDisplaycuda_types.cudnn9.cudnnBackendDescriptorType_t→format.CudaDisplaycuda_types.cudnn9.cudnnBackendNormMode_t→format.CudaDisplaycuda_types.cudnn9.cudnnBackendTensorReordering_t→format.CudaDisplaycuda_types.cudnn9.cudnnCTCGradMode_t→format.CudaDisplaycuda_types.cudnn9.cudnnDataType_t→format.CudaDisplaycuda_types.cudnn9.cudnnPointwiseMode_t→format.CudaDisplay
- Direction
- cuda_types.cudnn9 uses format, so a change to format can break cuda_types.cudnn9, while a change to cuda_types.cudnn9 can't break format.
- Position
- Below the diagonal: cuda_types.cudnn9 uses a module listed above it, which is the healthy direction.
14→9 cuda_types.cudnn9 uses cuda_types.cudnn✕
- Connections
- 9 pairs of types, where a type in cuda_types.cudnn9 uses a type in cuda_types.cudnn.
- The types involved
cuda_types.cudnn9.cudnnConvolutionBwdDataAlgoPerfStruct→cuda_types.cudnn.cudnnConvolutionBwdDataAlgo_tcuda_types.cudnn9.cudnnConvolutionBwdDataAlgoPerfStruct→cuda_types.cudnn.cudnnDeterminism_tcuda_types.cudnn9.cudnnConvolutionBwdDataAlgoPerfStruct→cuda_types.cudnn.cudnnMathType_tcuda_types.cudnn9.cudnnConvolutionBwdFilterAlgoPerfStruct→cuda_types.cudnn.cudnnConvolutionBwdFilterAlgo_tcuda_types.cudnn9.cudnnConvolutionBwdFilterAlgoPerfStruct→cuda_types.cudnn.cudnnDeterminism_tcuda_types.cudnn9.cudnnConvolutionBwdFilterAlgoPerfStruct→cuda_types.cudnn.cudnnMathType_tcuda_types.cudnn9.cudnnConvolutionFwdAlgoPerfStruct→cuda_types.cudnn.cudnnConvolutionFwdAlgo_tcuda_types.cudnn9.cudnnConvolutionFwdAlgoPerfStruct→cuda_types.cudnn.cudnnDeterminism_tcuda_types.cudnn9.cudnnConvolutionFwdAlgoPerfStruct→cuda_types.cudnn.cudnnMathType_t
- Direction
- cuda_types.cudnn9 uses cuda_types.cudnn, so a change to cuda_types.cudnn can break cuda_types.cudnn9, while a change to cuda_types.cudnn9 can't break cuda_types.cudnn.
- Position
- Below the diagonal: cuda_types.cudnn9 uses a module listed above it, which is the healthy direction.
14→20 cuda_types.cudnn9 uses miopen_sys part of a loop✕
- Connections
- 1 pair of types, where a type in cuda_types.cudnn9 uses a type in miopen_sys.
- The types involved
cuda_types.cudnn9.cudnnError_t→miopen_sys.miopenError_t
- Direction
- cuda_types.cudnn9 uses miopen_sys, so a change to miopen_sys can break cuda_types.cudnn9, while a change to cuda_types.cudnn9 can't break miopen_sys.
- Position
- Above the diagonal, so it is part of a loop. Neither module can be changed, tested or released on its own until one of these dependencies is removed.
15→3 cuda_types.cusparse uses zluda_common✕
- Connections
- 1 pair of types, where a type in cuda_types.cusparse uses a type in zluda_common.
- The types involved
cuda_types.cusparse.cusparseError_t→zluda_common.CudaErrorType
- Direction
- cuda_types.cusparse uses zluda_common, so a change to zluda_common can break cuda_types.cusparse, while a change to cuda_types.cusparse can't break zluda_common.
- Position
- Below the diagonal: cuda_types.cusparse uses a module listed above it, which is the healthy direction.
15→5 cuda_types.cusparse uses format✕
- Connections
- 26 pairs of types, where a type in cuda_types.cusparse uses a type in format. Showing 25 of them; the rest are in namespace-graph.json in this report's bundle.
- The types involved
cuda_types.cusparse.cusparseAction_t→format.CudaDisplaycuda_types.cusparse.cusparseColorAlg_t→format.CudaDisplaycuda_types.cusparse.cusparseCsr2CscAlg_t→format.CudaDisplaycuda_types.cusparse.cusparseDenseToSparseAlg_t→format.CudaDisplaycuda_types.cusparse.cusparseDiagType_t→format.CudaDisplaycuda_types.cusparse.cusparseDirection_t→format.CudaDisplaycuda_types.cusparse.cusparseFillMode_t→format.CudaDisplaycuda_types.cusparse.cusparseFormat_t→format.CudaDisplaycuda_types.cusparse.cusparseIndexBase_t→format.CudaDisplaycuda_types.cusparse.cusparseIndexType_t→format.CudaDisplaycuda_types.cusparse.cusparseMatrixType_t→format.CudaDisplaycuda_types.cusparse.cusparseOperation_t→format.CudaDisplaycuda_types.cusparse.cusparseOrder_t→format.CudaDisplaycuda_types.cusparse.cusparsePointerMode_t→format.CudaDisplaycuda_types.cusparse.cusparseSDDMMAlg_t→format.CudaDisplaycuda_types.cusparse.cusparseSolvePolicy_t→format.CudaDisplaycuda_types.cusparse.cusparseSpGEMMAlg_t→format.CudaDisplaycuda_types.cusparse.cusparseSpMMAlg_t→format.CudaDisplaycuda_types.cusparse.cusparseSpMMOpAlg_t→format.CudaDisplaycuda_types.cusparse.cusparseSpMVAlg_t→format.CudaDisplaycuda_types.cusparse.cusparseSpMatAttribute_t→format.CudaDisplaycuda_types.cusparse.cusparseSpSMAlg_t→format.CudaDisplaycuda_types.cusparse.cusparseSpSMUpdate_t→format.CudaDisplaycuda_types.cusparse.cusparseSpSVAlg_t→format.CudaDisplaycuda_types.cusparse.cusparseSpSVUpdate_t→format.CudaDisplay
- Direction
- cuda_types.cusparse uses format, so a change to format can break cuda_types.cusparse, while a change to cuda_types.cusparse can't break format.
- Position
- Below the diagonal: cuda_types.cusparse uses a module listed above it, which is the healthy direction.
15→12 cuda_types.cusparse uses dark_api✕
- Connections
- 4 pairs of types, where a type in cuda_types.cusparse uses a type in dark_api.
- The types involved
cuda_types.cusparse.cusparseDiagType_t→dark_api.ByteVecFficuda_types.cusparse.cusparseFillMode_t→dark_api.ByteVecFficuda_types.cusparse.cusparseIndexBase_t→dark_api.ByteVecFficuda_types.cusparse.cusparseMatrixType_t→dark_api.ByteVecFfi
- Direction
- cuda_types.cusparse uses dark_api, so a change to dark_api can break cuda_types.cusparse, while a change to cuda_types.cusparse can't break dark_api.
- Position
- Below the diagonal: cuda_types.cusparse uses a module listed above it, which is the healthy direction.
16→12 dark_api.cuda uses dark_api✕
- Connections
- 1 pair of types, where a type in dark_api.cuda uses a type in dark_api.
- The types involved
dark_api.cuda.CudaDarkApiGlobalTable→dark_api.DarkApiTable
- Direction
- dark_api.cuda uses dark_api, so a change to dark_api can break dark_api.cuda, while a change to dark_api.cuda can't break dark_api.
- Position
- Below the diagonal: dark_api.cuda uses a module listed above it, which is the healthy direction.
17→3 hipblaslt_sys uses zluda_common✕
- Connections
- 7 pairs of types, where a type in hipblaslt_sys uses a type in zluda_common.
- The types involved
hipblaslt_sys.hipDataType→zluda_common.FromCudahipblaslt_sys.hipblasComputeType_t→zluda_common.FromCudahipblaslt_sys.hipblasLtEpilogue_t→zluda_common.FromCudahipblaslt_sys.hipblasLtMatmulDescAttributes_t→zluda_common.FromCudahipblaslt_sys.hipblasLtMatmulPreferenceAttributes_t→zluda_common.FromCudahipblaslt_sys.hipblasLtMatrixLayoutAttribute_t→zluda_common.FromCudahipblaslt_sys.hipblasOperation_t→zluda_common.FromCuda
- Direction
- hipblaslt_sys uses zluda_common, so a change to zluda_common can break hipblaslt_sys, while a change to hipblaslt_sys can't break zluda_common.
- Position
- Below the diagonal: hipblaslt_sys uses a module listed above it, which is the healthy direction.
17→8 hipblaslt_sys uses cuda_types.cublaslt✕
- Connections
- 4 pairs of types, where a type in hipblaslt_sys uses a type in cuda_types.cublaslt.
- The types involved
hipblaslt_sys.hipblasLtEpilogue_t→cuda_types.cublaslt.cublasLtEpilogue_thipblaslt_sys.hipblasLtMatmulDescAttributes_t→cuda_types.cublaslt.cublasLtMatmulDescAttributes_thipblaslt_sys.hipblasLtMatmulPreferenceAttributes_t→cuda_types.cublaslt.cublasLtMatmulPreferenceAttributes_thipblaslt_sys.hipblasLtMatrixLayoutAttribute_t→cuda_types.cublaslt.cublasLtMatrixLayoutAttribute_t
- Direction
- hipblaslt_sys uses cuda_types.cublaslt, so a change to cuda_types.cublaslt can break hipblaslt_sys, while a change to hipblaslt_sys can't break cuda_types.cublaslt.
- Position
- Below the diagonal: hipblaslt_sys uses a module listed above it, which is the healthy direction.
17→19 hipblaslt_sys uses cuda_types.cublas part of a loop✕
- Connections
- 2 pairs of types, where a type in hipblaslt_sys uses a type in cuda_types.cublas.
- The types involved
hipblaslt_sys.hipblasComputeType_t→cuda_types.cublas.cublasComputeType_thipblaslt_sys.hipblasOperation_t→cuda_types.cublas.cublasOperation_t
- Direction
- hipblaslt_sys uses cuda_types.cublas, so a change to cuda_types.cublas can break hipblaslt_sys, while a change to hipblaslt_sys can't break cuda_types.cublas.
- Position
- Above the diagonal, so it is part of a loop. Neither module can be changed, tested or released on its own until one of these dependencies is removed.
18→7 ptx_parser_macros_impl.parser uses ptx_parser.ast✕
- Connections
- 1 pair of types, where a type in ptx_parser_macros_impl.parser uses a type in ptx_parser.ast.
- The types involved
ptx_parser_macros_impl.parser.Rule→ptx_parser.ast.Type
- Direction
- ptx_parser_macros_impl.parser uses ptx_parser.ast, so a change to ptx_parser.ast can break ptx_parser_macros_impl.parser, while a change to ptx_parser_macros_impl.parser can't break ptx_parser.ast.
- Position
- Below the diagonal: ptx_parser_macros_impl.parser uses a module listed above it, which is the healthy direction.
18→13 ptx_parser_macros_impl.parser uses ptx_parser_macros_impl✕
- Connections
- 1 pair of types, where a type in ptx_parser_macros_impl.parser uses a type in ptx_parser_macros_impl.
- The types involved
ptx_parser_macros_impl.parser.OpcodeDecl→ptx_parser_macros_impl.Arguments
- Direction
- ptx_parser_macros_impl.parser uses ptx_parser_macros_impl, so a change to ptx_parser_macros_impl can break ptx_parser_macros_impl.parser, while a change to ptx_parser_macros_impl.parser can't break ptx_parser_macros_impl.
- Position
- Below the diagonal: ptx_parser_macros_impl.parser uses a module listed above it, which is the healthy direction.
19→3 cuda_types.cublas uses zluda_common✕
- Connections
- 2 pairs of types, where a type in cuda_types.cublas uses a type in zluda_common.
- The types involved
cuda_types.cublas.cublasError_t→zluda_common.CudaErrorTypecuda_types.cublas.cublasMath_t→zluda_common.FromCuda
- Direction
- cuda_types.cublas uses zluda_common, so a change to zluda_common can break cuda_types.cublas, while a change to cuda_types.cublas can't break zluda_common.
- Position
- Below the diagonal: cuda_types.cublas uses a module listed above it, which is the healthy direction.
19→5 cuda_types.cublas uses format✕
- Connections
- 11 pairs of types, where a type in cuda_types.cublas uses a type in format.
- The types involved
cuda_types.cublas.cublasAtomicsMode_t→format.CudaDisplaycuda_types.cublas.cublasComputeType_t→format.CudaDisplaycuda_types.cublas.cublasDiagType_t→format.CudaDisplaycuda_types.cublas.cublasEmulationStrategy_t→format.CudaDisplaycuda_types.cublas.cublasFillMode_t→format.CudaDisplaycuda_types.cublas.cublasGemmAlgo_t→format.CudaDisplaycuda_types.cublas.cublasHandle_t→format.CudaDisplaycuda_types.cublas.cublasMath_t→format.CudaDisplaycuda_types.cublas.cublasOperation_t→format.CudaDisplaycuda_types.cublas.cublasPointerMode_t→format.CudaDisplaycuda_types.cublas.cublasSideMode_t→format.CudaDisplay
- Direction
- cuda_types.cublas uses format, so a change to format can break cuda_types.cublas, while a change to cuda_types.cublas can't break format.
- Position
- Below the diagonal: cuda_types.cublas uses a module listed above it, which is the healthy direction.
19→17 cuda_types.cublas uses hipblaslt_sys✕
- Connections
- 1 pair of types, where a type in cuda_types.cublas uses a type in hipblaslt_sys.
- The types involved
cuda_types.cublas.cublasError_t→hipblaslt_sys.hipblasLtError
- Direction
- cuda_types.cublas uses hipblaslt_sys, so a change to hipblaslt_sys can break cuda_types.cublas, while a change to cuda_types.cublas can't break hipblaslt_sys.
- Position
- Below the diagonal: cuda_types.cublas uses a module listed above it, which is the healthy direction.
20→3 miopen_sys uses zluda_common✕
- Connections
- 7 pairs of types, where a type in miopen_sys uses a type in zluda_common.
- The types involved
miopen_sys.miopenConvBwdDataAlgorithm_t→zluda_common.FromCudamiopen_sys.miopenConvBwdWeightsAlgorithm_t→zluda_common.FromCudamiopen_sys.miopenConvFwdAlgorithm_t→zluda_common.FromCudamiopen_sys.miopenConvolutionMode_t→zluda_common.FromCudamiopen_sys.miopenDataType_t→zluda_common.FromCudamiopen_sys.miopenError_t→zluda_common.CudaErrorTypemiopen_sys.miopenTensorLayout_t→zluda_common.FromCuda
- Direction
- miopen_sys uses zluda_common, so a change to zluda_common can break miopen_sys, while a change to miopen_sys can't break zluda_common.
- Position
- Below the diagonal: miopen_sys uses a module listed above it, which is the healthy direction.
20→9 miopen_sys uses cuda_types.cudnn✕
- Connections
- 5 pairs of types, where a type in miopen_sys uses a type in cuda_types.cudnn.
- The types involved
miopen_sys.miopenConvBwdDataAlgorithm_t→cuda_types.cudnn.cudnnConvolutionBwdDataAlgo_tmiopen_sys.miopenConvBwdWeightsAlgorithm_t→cuda_types.cudnn.cudnnConvolutionBwdFilterAlgo_tmiopen_sys.miopenConvFwdAlgorithm_t→cuda_types.cudnn.cudnnConvolutionFwdAlgo_tmiopen_sys.miopenConvolutionMode_t→cuda_types.cudnn.cudnnConvolutionMode_tmiopen_sys.miopenTensorLayout_t→cuda_types.cudnn.cudnnTensorFormat_t
- Direction
- miopen_sys uses cuda_types.cudnn, so a change to cuda_types.cudnn can break miopen_sys, while a change to miopen_sys can't break cuda_types.cudnn.
- Position
- Below the diagonal: miopen_sys uses a module listed above it, which is the healthy direction.
20→14 miopen_sys uses cuda_types.cudnn9✕
- Connections
- 1 pair of types, where a type in miopen_sys uses a type in cuda_types.cudnn9.
- The types involved
miopen_sys.miopenDataType_t→cuda_types.cudnn9.cudnnDataType_t
- Direction
- miopen_sys uses cuda_types.cudnn9, so a change to cuda_types.cudnn9 can break miopen_sys, while a change to miopen_sys can't break cuda_types.cudnn9.
- Position
- Below the diagonal: miopen_sys uses a module listed above it, which is the healthy direction.
21→2 ptx_parser uses llvm_zluda.utils✕
- Connections
- 1 pair of types, where a type in ptx_parser uses a type in llvm_zluda.utils.
- The types involved
ptx_parser.ptx_parser$module→llvm_zluda.utils.Module
- Direction
- ptx_parser uses llvm_zluda.utils, so a change to llvm_zluda.utils can break ptx_parser, while a change to ptx_parser can't break llvm_zluda.utils.
- Position
- Below the diagonal: ptx_parser uses a module listed above it, which is the healthy direction.
21→4 ptx_parser uses cuda_check.win✕
- Connections
- 1 pair of types, where a type in ptx_parser uses a type in cuda_check.win.
- The types involved
ptx_parser.TokenError→cuda_check.win.Error
- Direction
- ptx_parser uses cuda_check.win, so a change to cuda_check.win can break ptx_parser, while a change to ptx_parser can't break cuda_check.win.
- Position
- Below the diagonal: ptx_parser uses a module listed above it, which is the healthy direction.
21→7 ptx_parser uses ptx_parser.ast✕
- Connections
- 4 pairs of types, where a type in ptx_parser uses a type in ptx_parser.ast.
- The types involved
ptx_parser.PtxParserState→ptx_parser.ast.MethodDeclarationptx_parser.ScalarType→ptx_parser.ast.ScalarKindptx_parser.ptx_parser$module→ptx_parser.ast.ImmediateValueptx_parser.ptx_parser$module→ptx_parser.ast.RegOrImmediate
- Direction
- ptx_parser uses ptx_parser.ast, so a change to ptx_parser.ast can break ptx_parser, while a change to ptx_parser can't break ptx_parser.ast.
- Position
- Below the diagonal: ptx_parser uses a module listed above it, which is the healthy direction.
21→18 ptx_parser uses ptx_parser_macros_impl.parser✕
- Connections
- 1 pair of types, where a type in ptx_parser uses a type in ptx_parser_macros_impl.parser.
- The types involved
ptx_parser.ptx_parser$module→ptx_parser_macros_impl.parser.Instruction
- Direction
- ptx_parser uses ptx_parser_macros_impl.parser, so a change to ptx_parser_macros_impl.parser can break ptx_parser, while a change to ptx_parser can't break ptx_parser_macros_impl.parser.
- Position
- Below the diagonal: ptx_parser uses a module listed above it, which is the healthy direction.
22→5 zluda_trace uses format✕
- Connections
- 1 pair of types, where a type in zluda_trace uses a type in format.
- The types involved
zluda_trace.CudaResult→format.CudaDisplay
- Direction
- zluda_trace uses format, so a change to format can break zluda_trace, while a change to zluda_trace can't break format.
- Position
- Below the diagonal: zluda_trace uses a module listed above it, which is the healthy direction.
22→16 zluda_trace uses dark_api.cuda✕
- Connections
- 2 pairs of types, where a type in zluda_trace uses a type in dark_api.cuda.
- The types involved
zluda_trace.DarkApiTrace→dark_api.cuda.CudaDarkApizluda_trace.InternalTableImpl→dark_api.cuda.CudaDarkApi
- Direction
- zluda_trace uses dark_api.cuda, so a change to dark_api.cuda can break zluda_trace, while a change to zluda_trace can't break dark_api.cuda.
- Position
- Below the diagonal: zluda_trace uses a module listed above it, which is the healthy direction.
22→26 zluda_trace uses zluda_trace.log part of a loop✕
- Connections
- 6 pairs of types, where a type in zluda_trace uses a type in zluda_trace.log.
- The types involved
zluda_trace.FnCallLog→zluda_trace.log.CudaFunctionNamezluda_trace.FnCallLog→zluda_trace.log.ErrorEntryzluda_trace.GlobalDelayedState→zluda_trace.log.Writerzluda_trace.GlobalState2→zluda_trace.log.CudaFunctionNamezluda_trace.GlobalState2→zluda_trace.log.Writerzluda_trace.OuterCallGuard→zluda_trace.log.Writer
- Direction
- zluda_trace uses zluda_trace.log, so a change to zluda_trace.log can break zluda_trace, while a change to zluda_trace can't break zluda_trace.log.
- Position
- Above the diagonal, so it is part of a loop. Neither module can be changed, tested or released on its own until one of these dependencies is removed.
22→29 zluda_trace uses cuda_types.cuda part of a loop✕
- Connections
- 4 pairs of types, where a type in zluda_trace uses a type in cuda_types.cuda.
- The types involved
zluda_trace.zluda_trace$module→cuda_types.cuda.CUfunctionzluda_trace.zluda_trace$module→cuda_types.cuda.CUmodulezluda_trace.zluda_trace$module→cuda_types.cuda.CUstreamzluda_trace.zluda_trace$module→cuda_types.cuda.CUuuid_st
- Direction
- zluda_trace uses cuda_types.cuda, so a change to cuda_types.cuda can break zluda_trace, while a change to zluda_trace can't break cuda_types.cuda.
- Position
- Above the diagonal, so it is part of a loop. Neither module can be changed, tested or released on its own until one of these dependencies is removed.
23→3 cuda_types.cudnn8 uses zluda_common✕
- Connections
- 1 pair of types, where a type in cuda_types.cudnn8 uses a type in zluda_common.
- The types involved
cuda_types.cudnn8.cudnnError_t→zluda_common.CudaErrorType
- Direction
- cuda_types.cudnn8 uses zluda_common, so a change to zluda_common can break cuda_types.cudnn8, while a change to cuda_types.cudnn8 can't break zluda_common.
- Position
- Below the diagonal: cuda_types.cudnn8 uses a module listed above it, which is the healthy direction.
23→5 cuda_types.cudnn8 uses format✕
- Connections
- 8 pairs of types, where a type in cuda_types.cudnn8 uses a type in format.
- The types involved
cuda_types.cudnn8.cudnnAlgorithmUnionStruct→format.CudaDisplaycuda_types.cudnn8.cudnnBackendBehaviorNote_t→format.CudaDisplaycuda_types.cudnn8.cudnnBackendKnobType_t→format.CudaDisplaycuda_types.cudnn8.cudnnBackendNumericalNote_t→format.CudaDisplaycuda_types.cudnn8.cudnnConvolutionBwdDataAlgoPerfStruct→format.CudaDisplaycuda_types.cudnn8.cudnnConvolutionBwdFilterAlgoPerfStruct→format.CudaDisplaycuda_types.cudnn8.cudnnConvolutionFwdAlgoPerfStruct→format.CudaDisplaycuda_types.cudnn8.cudnnDebugStruct→format.CudaDisplay
- Direction
- cuda_types.cudnn8 uses format, so a change to format can break cuda_types.cudnn8, while a change to cuda_types.cudnn8 can't break format.
- Position
- Below the diagonal: cuda_types.cudnn8 uses a module listed above it, which is the healthy direction.
23→9 cuda_types.cudnn8 uses cuda_types.cudnn✕
- Connections
- 9 pairs of types, where a type in cuda_types.cudnn8 uses a type in cuda_types.cudnn.
- The types involved
cuda_types.cudnn8.cudnnConvolutionBwdDataAlgoPerfStruct→cuda_types.cudnn.cudnnConvolutionBwdDataAlgo_tcuda_types.cudnn8.cudnnConvolutionBwdDataAlgoPerfStruct→cuda_types.cudnn.cudnnDeterminism_tcuda_types.cudnn8.cudnnConvolutionBwdDataAlgoPerfStruct→cuda_types.cudnn.cudnnMathType_tcuda_types.cudnn8.cudnnConvolutionBwdFilterAlgoPerfStruct→cuda_types.cudnn.cudnnConvolutionBwdFilterAlgo_tcuda_types.cudnn8.cudnnConvolutionBwdFilterAlgoPerfStruct→cuda_types.cudnn.cudnnDeterminism_tcuda_types.cudnn8.cudnnConvolutionBwdFilterAlgoPerfStruct→cuda_types.cudnn.cudnnMathType_tcuda_types.cudnn8.cudnnConvolutionFwdAlgoPerfStruct→cuda_types.cudnn.cudnnConvolutionFwdAlgo_tcuda_types.cudnn8.cudnnConvolutionFwdAlgoPerfStruct→cuda_types.cudnn.cudnnDeterminism_tcuda_types.cudnn8.cudnnConvolutionFwdAlgoPerfStruct→cuda_types.cudnn.cudnnMathType_t
- Direction
- cuda_types.cudnn8 uses cuda_types.cudnn, so a change to cuda_types.cudnn can break cuda_types.cudnn8, while a change to cuda_types.cudnn8 can't break cuda_types.cudnn.
- Position
- Below the diagonal: cuda_types.cudnn8 uses a module listed above it, which is the healthy direction.
23→20 cuda_types.cudnn8 uses miopen_sys✕
- Connections
- 1 pair of types, where a type in cuda_types.cudnn8 uses a type in miopen_sys.
- The types involved
cuda_types.cudnn8.cudnnError_t→miopen_sys.miopenError_t
- Direction
- cuda_types.cudnn8 uses miopen_sys, so a change to miopen_sys can break cuda_types.cudnn8, while a change to cuda_types.cudnn8 can't break miopen_sys.
- Position
- Below the diagonal: cuda_types.cudnn8 uses a module listed above it, which is the healthy direction.
24→7 ptx.pass.instruction_mode_to_global_mode uses ptx_parser.ast✕
- Connections
- 3 pairs of types, where a type in ptx.pass.instruction_mode_to_global_mode uses a type in ptx_parser.ast.
- The types involved
ptx.pass.instruction_mode_to_global_mode.InstructionModes→ptx_parser.ast.ArithFloatptx.pass.instruction_mode_to_global_mode.InstructionModes→ptx_parser.ast.CvtDetailsptx.pass.instruction_mode_to_global_mode.InstructionModes→ptx_parser.ast.RcpData
- Direction
- ptx.pass.instruction_mode_to_global_mode uses ptx_parser.ast, so a change to ptx_parser.ast can break ptx.pass.instruction_mode_to_global_mode, while a change to ptx.pass.instruction_mode_to_global_mode can't break ptx_parser.ast.
- Position
- Below the diagonal: ptx.pass.instruction_mode_to_global_mode uses a module listed above it, which is the healthy direction.
24→18 ptx.pass.instruction_mode_to_global_mode uses ptx_parser_macros_impl.parser✕
- Connections
- 1 pair of types, where a type in ptx.pass.instruction_mode_to_global_mode uses a type in ptx_parser_macros_impl.parser.
- The types involved
ptx.pass.instruction_mode_to_global_mode.instruction_mode_to_global_mode$module→ptx_parser_macros_impl.parser.Instruction
- Direction
- ptx.pass.instruction_mode_to_global_mode uses ptx_parser_macros_impl.parser, so a change to ptx_parser_macros_impl.parser can break ptx.pass.instruction_mode_to_global_mode, while a change to ptx.pass.instruction_mode_to_global_mode can't break ptx_parser_macros_impl.parser.
- Position
- Below the diagonal: ptx.pass.instruction_mode_to_global_mode uses a module listed above it, which is the healthy direction.
24→21 ptx.pass.instruction_mode_to_global_mode uses ptx_parser✕
- Connections
- 1 pair of types, where a type in ptx.pass.instruction_mode_to_global_mode uses a type in ptx_parser.
- The types involved
ptx.pass.instruction_mode_to_global_mode.InstructionModes→ptx_parser.ScalarType
- Direction
- ptx.pass.instruction_mode_to_global_mode uses ptx_parser, so a change to ptx_parser can break ptx.pass.instruction_mode_to_global_mode, while a change to ptx.pass.instruction_mode_to_global_mode can't break ptx_parser.
- Position
- Below the diagonal: ptx.pass.instruction_mode_to_global_mode uses a module listed above it, which is the healthy direction.
24→30 ptx.pass.instruction_mode_to_global_mode uses ptx.pass part of a loop✕
- Connections
- 19 pairs of types, where a type in ptx.pass.instruction_mode_to_global_mode uses a type in ptx.pass.
- The types involved
ptx.pass.instruction_mode_to_global_mode.BasicBlockControlState→ptx.pass.SpirvWordptx.pass.instruction_mode_to_global_mode.BasicBlockControlState→ptx.pass.Statementptx.pass.instruction_mode_to_global_mode.BasicBlockEntryState→ptx.pass.SpirvWordptx.pass.instruction_mode_to_global_mode.BasicBlockState→ptx.pass.SpirvWordptx.pass.instruction_mode_to_global_mode.BasicBlockState→ptx.pass.Statementptx.pass.instruction_mode_to_global_mode.ControlFlowGraph→ptx.pass.SpirvWordptx.pass.instruction_mode_to_global_mode.DenormalF16F64View→ptx.pass.ModeRegisterptx.pass.instruction_mode_to_global_mode.DenormalF32View→ptx.pass.ModeRegisterptx.pass.instruction_mode_to_global_mode.FullBasicBlockEntryState→ptx.pass.SpirvWordptx.pass.instruction_mode_to_global_mode.Mode→ptx.pass.SpirvWordptx.pass.instruction_mode_to_global_mode.ModeView→ptx.pass.ModeRegisterptx.pass.instruction_mode_to_global_mode.Node→ptx.pass.SpirvWordptx.pass.instruction_mode_to_global_mode.ResolvedNode→ptx.pass.SpirvWordptx.pass.instruction_mode_to_global_mode.RoundingF16F64View→ptx.pass.ModeRegisterptx.pass.instruction_mode_to_global_mode.RoundingF32View→ptx.pass.ModeRegisterptx.pass.instruction_mode_to_global_mode.instruction_mode_to_global_mode$module→ptx.pass.Directive2ptx.pass.instruction_mode_to_global_mode.instruction_mode_to_global_mode$module→ptx.pass.Functionptx.pass.instruction_mode_to_global_mode.instruction_mode_to_global_mode$module→ptx.pass.GlobalStringIdentResolver2ptx.pass.instruction_mode_to_global_mode.instruction_mode_to_global_mode$module→ptx.pass.SpirvWord
- Direction
- ptx.pass.instruction_mode_to_global_mode uses ptx.pass, so a change to ptx.pass can break ptx.pass.instruction_mode_to_global_mode, while a change to ptx.pass.instruction_mode_to_global_mode can't break ptx.pass.
- Position
- Above the diagonal, so it is part of a loop. Neither module can be changed, tested or released on its own until one of these dependencies is removed.
25→2 zluda_dnn.impl uses llvm_zluda.utils✕
- Connections
- 1 pair of types, where a type in zluda_dnn.impl uses a type in llvm_zluda.utils.
- The types involved
zluda_dnn.impl.impl$module→llvm_zluda.utils.Context
- Direction
- zluda_dnn.impl uses llvm_zluda.utils, so a change to llvm_zluda.utils can break zluda_dnn.impl, while a change to zluda_dnn.impl can't break llvm_zluda.utils.
- Position
- Below the diagonal: zluda_dnn.impl uses a module listed above it, which is the healthy direction.
25→3 zluda_dnn.impl uses zluda_common✕
- Connections
- 2 pairs of types, where a type in zluda_dnn.impl uses a type in zluda_common.
- The types involved
zluda_dnn.impl.Context→zluda_common.FromCudazluda_dnn.impl.Context→zluda_common.ZludaObject
- Direction
- zluda_dnn.impl uses zluda_common, so a change to zluda_common can break zluda_dnn.impl, while a change to zluda_dnn.impl can't break zluda_common.
- Position
- Below the diagonal: zluda_dnn.impl uses a module listed above it, which is the healthy direction.
25→6 zluda_dnn.impl uses hip_runtime_sys✕
- Connections
- 2 pairs of types, where a type in zluda_dnn.impl uses a type in hip_runtime_sys.
- The types involved
zluda_dnn.impl.TemporaryBufferAllocator→hip_runtime_sys.hipStream_tzluda_dnn.impl.impl$module→hip_runtime_sys.hipStream_t
- Direction
- zluda_dnn.impl uses hip_runtime_sys, so a change to hip_runtime_sys can break zluda_dnn.impl, while a change to zluda_dnn.impl can't break hip_runtime_sys.
- Position
- Below the diagonal: zluda_dnn.impl uses a module listed above it, which is the healthy direction.
25→9 zluda_dnn.impl uses cuda_types.cudnn✕
- Connections
- 5 pairs of types, where a type in zluda_dnn.impl uses a type in cuda_types.cudnn.
- The types involved
zluda_dnn.impl.impl$module→cuda_types.cudnn.cudnnConvolutionBwdDataAlgo_tzluda_dnn.impl.impl$module→cuda_types.cudnn.cudnnConvolutionBwdFilterAlgo_tzluda_dnn.impl.impl$module→cuda_types.cudnn.cudnnConvolutionFwdAlgo_tzluda_dnn.impl.impl$module→cuda_types.cudnn.cudnnMathType_tzluda_dnn.impl.impl$module→cuda_types.cudnn.cudnnTensorFormat_t
- Direction
- zluda_dnn.impl uses cuda_types.cudnn, so a change to cuda_types.cudnn can break zluda_dnn.impl, while a change to zluda_dnn.impl can't break cuda_types.cudnn.
- Position
- Below the diagonal: zluda_dnn.impl uses a module listed above it, which is the healthy direction.
25→14 zluda_dnn.impl uses cuda_types.cudnn9✕
- Connections
- 1 pair of types, where a type in zluda_dnn.impl uses a type in cuda_types.cudnn9.
- The types involved
zluda_dnn.impl.impl$module→cuda_types.cudnn9.cudnnDataType_t
- Direction
- zluda_dnn.impl uses cuda_types.cudnn9, so a change to cuda_types.cudnn9 can break zluda_dnn.impl, while a change to zluda_dnn.impl can't break cuda_types.cudnn9.
- Position
- Below the diagonal: zluda_dnn.impl uses a module listed above it, which is the healthy direction.
25→20 zluda_dnn.impl uses miopen_sys✕
- Connections
- 25 pairs of types, where a type in zluda_dnn.impl uses a type in miopen_sys.
- The types involved
zluda_dnn.impl.Context→miopen_sys.miopenHandle_tzluda_dnn.impl.ConvolutionDescriptorCacheKey→miopen_sys.miopenConvolutionDescriptor_tzluda_dnn.impl.ConvolutionDescriptorCacheKey→miopen_sys.miopenConvolutionMode_tzluda_dnn.impl.ConvolutionOpCacheKey→miopen_sys.miopenConvolutionDescriptor_tzluda_dnn.impl.ConvolutionOpCacheKey→miopen_sys.miopenProblemDirection_tzluda_dnn.impl.ConvolutionOpCacheKey→miopen_sys.miopenTensorDescriptor_tzluda_dnn.impl.SearchCache→miopen_sys.miopenConvolutionDescriptor_tzluda_dnn.impl.SearchCache→miopen_sys.miopenHandle_tzluda_dnn.impl.SearchCache→miopen_sys.miopenProblemDirection_tzluda_dnn.impl.SearchCache→miopen_sys.miopenTensorDescriptor_tzluda_dnn.impl.SolutionWithProperties→miopen_sys.miopenConvAlgorithm_tzluda_dnn.impl.SolutionWithProperties→miopen_sys.miopenSolution_tzluda_dnn.impl.TensorCacheKey→miopen_sys.miopenDataType_tzluda_dnn.impl.TensorCacheKey→miopen_sys.miopenTensorDescriptor_tzluda_dnn.impl.impl$module→miopen_sys.miopenConvAlgorithm_tzluda_dnn.impl.impl$module→miopen_sys.miopenConvBwdDataAlgorithm_tzluda_dnn.impl.impl$module→miopen_sys.miopenConvBwdWeightsAlgorithm_tzluda_dnn.impl.impl$module→miopen_sys.miopenConvFwdAlgorithm_tzluda_dnn.impl.impl$module→miopen_sys.miopenConvolutionDescriptor_tzluda_dnn.impl.impl$module→miopen_sys.miopenConvolutionMode_tzluda_dnn.impl.impl$module→miopen_sys.miopenDataType_tzluda_dnn.impl.impl$module→miopen_sys.miopenDropoutDescriptor_tzluda_dnn.impl.impl$module→miopen_sys.miopenProblemDirection_tzluda_dnn.impl.impl$module→miopen_sys.miopenTensorDescriptor_tzluda_dnn.impl.impl$module→miopen_sys.miopenTensorLayout_t
- Direction
- zluda_dnn.impl uses miopen_sys, so a change to miopen_sys can break zluda_dnn.impl, while a change to zluda_dnn.impl can't break miopen_sys.
- Position
- Below the diagonal: zluda_dnn.impl uses a module listed above it, which is the healthy direction.
26→22 zluda_trace.log uses zluda_trace✕
- Connections
- 2 pairs of types, where a type in zluda_trace.log uses a type in zluda_trace.
- The types involved
zluda_trace.log.Writer→zluda_trace.FnCallLogzluda_trace.log.Writer→zluda_trace.Settings
- Direction
- zluda_trace.log uses zluda_trace, so a change to zluda_trace can break zluda_trace.log, while a change to zluda_trace.log can't break zluda_trace.
- Position
- Below the diagonal: zluda_trace.log uses a module listed above it, which is the healthy direction.
27→2 ptx.pass.llvm.emit uses llvm_zluda.utils✕
- Connections
- 4 pairs of types, where a type in ptx.pass.llvm.emit uses a type in llvm_zluda.utils.
- The types involved
ptx.pass.llvm.emit.Builder→llvm_zluda.utils.Contextptx.pass.llvm.emit.ModuleEmitContext→llvm_zluda.utils.Contextptx.pass.llvm.emit.ModuleEmitContext→llvm_zluda.utils.Moduleptx.pass.llvm.emit.emit$module→llvm_zluda.utils.Context
- Direction
- ptx.pass.llvm.emit uses llvm_zluda.utils, so a change to llvm_zluda.utils can break ptx.pass.llvm.emit, while a change to ptx.pass.llvm.emit can't break llvm_zluda.utils.
- Position
- Below the diagonal: ptx.pass.llvm.emit uses a module listed above it, which is the healthy direction.
27→7 ptx.pass.llvm.emit uses ptx_parser.ast✕
- Connections
- 44 pairs of types, where a type in ptx.pass.llvm.emit uses a type in ptx_parser.ast. Showing 25 of them; the rest are in namespace-graph.json in this report's bundle.
- The types involved
ptx.pass.llvm.emit.MethodEmitContext→ptx_parser.ast.ArithDetailsptx.pass.llvm.emit.MethodEmitContext→ptx_parser.ast.ArithFloatptx.pass.llvm.emit.MethodEmitContext→ptx_parser.ast.ArithIntegerptx.pass.llvm.emit.MethodEmitContext→ptx_parser.ast.AtomCasDetailsptx.pass.llvm.emit.MethodEmitContext→ptx_parser.ast.AtomDetailsptx.pass.llvm.emit.MethodEmitContext→ptx_parser.ast.CallArgsptx.pass.llvm.emit.MethodEmitContext→ptx_parser.ast.CallDetailsptx.pass.llvm.emit.MethodEmitContext→ptx_parser.ast.CarryDetailsptx.pass.llvm.emit.MethodEmitContext→ptx_parser.ast.CpAsyncDetailsptx.pass.llvm.emit.MethodEmitContext→ptx_parser.ast.CreatePolicyFractionalDetailsptx.pass.llvm.emit.MethodEmitContext→ptx_parser.ast.CvtDetailsptx.pass.llvm.emit.MethodEmitContext→ptx_parser.ast.CvtaDetailsptx.pass.llvm.emit.MethodEmitContext→ptx_parser.ast.DivDetailsptx.pass.llvm.emit.MethodEmitContext→ptx_parser.ast.DivFloatDetailsptx.pass.llvm.emit.MethodEmitContext→ptx_parser.ast.Dp4aDetailsptx.pass.llvm.emit.MethodEmitContext→ptx_parser.ast.FlushToZeroptx.pass.llvm.emit.MethodEmitContext→ptx_parser.ast.LdDetailsptx.pass.llvm.emit.MethodEmitContext→ptx_parser.ast.MadCarryDetailsptx.pass.llvm.emit.MethodEmitContext→ptx_parser.ast.MadDetailsptx.pass.llvm.emit.MethodEmitContext→ptx_parser.ast.MinMaxDetailsptx.pass.llvm.emit.MethodEmitContext→ptx_parser.ast.Mul24Detailsptx.pass.llvm.emit.MethodEmitContext→ptx_parser.ast.MulDetailsptx.pass.llvm.emit.MethodEmitContext→ptx_parser.ast.RcpDataptx.pass.llvm.emit.MethodEmitContext→ptx_parser.ast.RetDataptx.pass.llvm.emit.MethodEmitContext→ptx_parser.ast.SetBoolData
- Direction
- ptx.pass.llvm.emit uses ptx_parser.ast, so a change to ptx_parser.ast can break ptx.pass.llvm.emit, while a change to ptx.pass.llvm.emit can't break ptx_parser.ast.
- Position
- Below the diagonal: ptx.pass.llvm.emit uses a module listed above it, which is the healthy direction.
27→18 ptx.pass.llvm.emit uses ptx_parser_macros_impl.parser✕
- Connections
- 1 pair of types, where a type in ptx.pass.llvm.emit uses a type in ptx_parser_macros_impl.parser.
- The types involved
ptx.pass.llvm.emit.MethodEmitContext→ptx_parser_macros_impl.parser.Instruction
- Direction
- ptx.pass.llvm.emit uses ptx_parser_macros_impl.parser, so a change to ptx_parser_macros_impl.parser can break ptx.pass.llvm.emit, while a change to ptx.pass.llvm.emit can't break ptx_parser_macros_impl.parser.
- Position
- Below the diagonal: ptx.pass.llvm.emit uses a module listed above it, which is the healthy direction.
27→21 ptx.pass.llvm.emit uses ptx_parser✕
- Connections
- 9 pairs of types, where a type in ptx.pass.llvm.emit uses a type in ptx_parser.
- The types involved
ptx.pass.llvm.emit.FloatingPoint→ptx_parser.ScalarTypeptx.pass.llvm.emit.LLVMTypeDisplay→ptx_parser.ScalarTypeptx.pass.llvm.emit.MethodEmitContext→ptx_parser.MemScopeptx.pass.llvm.emit.MethodEmitContext→ptx_parser.ScalarTypeptx.pass.llvm.emit.MethodEmitContext→ptx_parser.StateSpaceptx.pass.llvm.emit.emit$module→ptx_parser.AtomSemanticsptx.pass.llvm.emit.emit$module→ptx_parser.MemScopeptx.pass.llvm.emit.emit$module→ptx_parser.ScalarTypeptx.pass.llvm.emit.emit$module→ptx_parser.StateSpace
- Direction
- ptx.pass.llvm.emit uses ptx_parser, so a change to ptx_parser can break ptx.pass.llvm.emit, while a change to ptx.pass.llvm.emit can't break ptx_parser.
- Position
- Below the diagonal: ptx.pass.llvm.emit uses a module listed above it, which is the healthy direction.
27→24 ptx.pass.llvm.emit uses ptx.pass.instruction_mode_to_global_mode✕
- Connections
- 1 pair of types, where a type in ptx.pass.llvm.emit uses a type in ptx.pass.instruction_mode_to_global_mode.
- The types involved
ptx.pass.llvm.emit.MethodEmitContext→ptx.pass.instruction_mode_to_global_mode.RoundingMode
- Direction
- ptx.pass.llvm.emit uses ptx.pass.instruction_mode_to_global_mode, so a change to ptx.pass.instruction_mode_to_global_mode can break ptx.pass.llvm.emit, while a change to ptx.pass.llvm.emit can't break ptx.pass.instruction_mode_to_global_mode.
- Position
- Below the diagonal: ptx.pass.llvm.emit uses a module listed above it, which is the healthy direction.
27→30 ptx.pass.llvm.emit uses ptx.pass part of a loop✕
- Connections
- 19 pairs of types, where a type in ptx.pass.llvm.emit uses a type in ptx.pass.
- The types involved
ptx.pass.llvm.emit.FloatingPoint→ptx.pass.SpirvWordptx.pass.llvm.emit.MethodEmitContext→ptx.pass.BranchConditionptx.pass.llvm.emit.MethodEmitContext→ptx.pass.ConstantDefinitionptx.pass.llvm.emit.MethodEmitContext→ptx.pass.ImplicitConversionptx.pass.llvm.emit.MethodEmitContext→ptx.pass.KernelAttributesptx.pass.llvm.emit.MethodEmitContext→ptx.pass.ModeRegisterptx.pass.llvm.emit.MethodEmitContext→ptx.pass.PtrAccessptx.pass.llvm.emit.MethodEmitContext→ptx.pass.RepackVectorDetailsptx.pass.llvm.emit.MethodEmitContext→ptx.pass.SpirvWordptx.pass.llvm.emit.MethodEmitContext→ptx.pass.Statementptx.pass.llvm.emit.MethodEmitContext→ptx.pass.VectorReadptx.pass.llvm.emit.MethodEmitContext→ptx.pass.VectorWriteptx.pass.llvm.emit.ModuleEmitContext→ptx.pass.Functionptx.pass.llvm.emit.ModuleEmitContext→ptx.pass.GlobalStringIdentResolver2ptx.pass.llvm.emit.ResolveIdent→ptx.pass.GlobalStringIdentResolver2ptx.pass.llvm.emit.ResolveIdent→ptx.pass.SpirvWordptx.pass.llvm.emit.emit$module→ptx.pass.Directive2ptx.pass.llvm.emit.emit$module→ptx.pass.GlobalStringIdentResolver2ptx.pass.llvm.emit.emit$module→ptx.pass.SpirvWord
- Direction
- ptx.pass.llvm.emit uses ptx.pass, so a change to ptx.pass can break ptx.pass.llvm.emit, while a change to ptx.pass.llvm.emit can't break ptx.pass.
- Position
- Above the diagonal, so it is part of a loop. Neither module can be changed, tested or released on its own until one of these dependencies is removed.
28→12 zluda_server_common uses dark_api✕
- Connections
- 1 pair of types, where a type in zluda_server_common uses a type in dark_api.
- The types involved
zluda_server_common.zludaGetFunctionArgsOut→dark_api.FunctionArgInfo
- Direction
- zluda_server_common uses dark_api, so a change to dark_api can break zluda_server_common, while a change to zluda_server_common can't break dark_api.
- Position
- Below the diagonal: zluda_server_common uses a module listed above it, which is the healthy direction.
28→26 zluda_server_common uses zluda_trace.log✕
- Connections
- 1 pair of types, where a type in zluda_server_common uses a type in zluda_trace.log.
- The types involved
zluda_server_common.SliceWriter→zluda_trace.log.Writer
- Direction
- zluda_server_common uses zluda_trace.log, so a change to zluda_trace.log can break zluda_server_common, while a change to zluda_server_common can't break zluda_trace.log.
- Position
- Below the diagonal: zluda_server_common uses a module listed above it, which is the healthy direction.
29→3 cuda_types.cuda uses zluda_common✕
- Connections
- 3 pairs of types, where a type in cuda_types.cuda uses a type in zluda_common.
- The types involved
cuda_types.cuda.CUerror→zluda_common.CudaErrorTypecuda_types.cuda.ID3D11Resource→zluda_common.FromCudacuda_types.cuda.IDXGIAdapter→zluda_common.FromCuda
- Direction
- cuda_types.cuda uses zluda_common, so a change to zluda_common can break cuda_types.cuda, while a change to cuda_types.cuda can't break zluda_common.
- Position
- Below the diagonal: cuda_types.cuda uses a module listed above it, which is the healthy direction.
29→5 cuda_types.cuda uses format✕
- Connections
- 238 pairs of types, where a type in cuda_types.cuda uses a type in format. Showing 25 of them; the rest are in namespace-graph.json in this report's bundle.
- The types involved
cuda_types.cuda.CUCoredumpGenerationFlags→format.CudaDisplaycuda_types.cuda.CUD3D10map_flags_enum→format.CudaDisplaycuda_types.cuda.CUD3D10register_flags_enum→format.CudaDisplaycuda_types.cuda.CUDA_ARRAY3D_DESCRIPTOR_st→format.CudaDisplaycuda_types.cuda.CUDA_ARRAY3D_DESCRIPTOR_v1_st→format.CudaDisplaycuda_types.cuda.CUDA_ARRAY_DESCRIPTOR_st→format.CudaDisplaycuda_types.cuda.CUDA_ARRAY_DESCRIPTOR_v1_st→format.CudaDisplaycuda_types.cuda.CUDA_ARRAY_MEMORY_REQUIREMENTS_st→format.CudaDisplaycuda_types.cuda.CUDA_ARRAY_SPARSE_PROPERTIES_st→format.CudaDisplaycuda_types.cuda.CUDA_ARRAY_SPARSE_PROPERTIES_st__bindgen_ty_1→format.CudaDisplaycuda_types.cuda.CUDA_BATCH_MEM_OP_NODE_PARAMS_v1_st→format.CudaDisplaycuda_types.cuda.CUDA_BATCH_MEM_OP_NODE_PARAMS_v2_st→format.CudaDisplaycuda_types.cuda.CUDA_CHILD_GRAPH_NODE_PARAMS_st→format.CudaDisplaycuda_types.cuda.CUDA_CONDITIONAL_NODE_PARAMS→format.CudaDisplaycuda_types.cuda.CUDA_EVENT_RECORD_NODE_PARAMS_st→format.CudaDisplaycuda_types.cuda.CUDA_EVENT_WAIT_NODE_PARAMS_st→format.CudaDisplaycuda_types.cuda.CUDA_EXTERNAL_MEMORY_BUFFER_DESC_st→format.CudaDisplaycuda_types.cuda.CUDA_EXTERNAL_MEMORY_HANDLE_DESC_st→format.CudaDisplaycuda_types.cuda.CUDA_EXTERNAL_MEMORY_HANDLE_DESC_st__bindgen_ty_1__bindgen_ty_1→format.CudaDisplaycuda_types.cuda.CUDA_EXTERNAL_MEMORY_MIPMAPPED_ARRAY_DESC_st→format.CudaDisplaycuda_types.cuda.CUDA_EXTERNAL_SEMAPHORE_HANDLE_DESC_st→format.CudaDisplaycuda_types.cuda.CUDA_EXTERNAL_SEMAPHORE_HANDLE_DESC_st__bindgen_ty_1__bindgen_ty_1→format.CudaDisplaycuda_types.cuda.CUDA_EXTERNAL_SEMAPHORE_SIGNAL_PARAMS_st→format.CudaDisplaycuda_types.cuda.CUDA_EXTERNAL_SEMAPHORE_SIGNAL_PARAMS_st__bindgen_ty_1→format.CudaDisplaycuda_types.cuda.CUDA_EXTERNAL_SEMAPHORE_SIGNAL_PARAMS_st__bindgen_ty_1__bindgen_ty_1→format.CudaDisplay
- Direction
- cuda_types.cuda uses format, so a change to format can break cuda_types.cuda, while a change to cuda_types.cuda can't break format.
- Position
- Below the diagonal: cuda_types.cuda uses a module listed above it, which is the healthy direction.
29→6 cuda_types.cuda uses hip_runtime_sys✕
- Connections
- 1 pair of types, where a type in cuda_types.cuda uses a type in hip_runtime_sys.
- The types involved
cuda_types.cuda.CUerror→hip_runtime_sys.hipErrorCode_t
- Direction
- cuda_types.cuda uses hip_runtime_sys, so a change to hip_runtime_sys can break cuda_types.cuda, while a change to cuda_types.cuda can't break hip_runtime_sys.
- Position
- Below the diagonal: cuda_types.cuda uses a module listed above it, which is the healthy direction.
29→28 cuda_types.cuda uses zluda_server_common✕
- Connections
- 5 pairs of types, where a type in cuda_types.cuda uses a type in zluda_server_common.
- The types involved
cuda_types.cuda.CUcontext→zluda_server_common.CudaEncodecuda_types.cuda.CUdeviceptr_v2→zluda_server_common.CudaEncodecuda_types.cuda.CUfunction→zluda_server_common.CudaEncodecuda_types.cuda.CUmodule→zluda_server_common.CudaEncodecuda_types.cuda.CUstream→zluda_server_common.CudaEncode
- Direction
- cuda_types.cuda uses zluda_server_common, so a change to zluda_server_common can break cuda_types.cuda, while a change to cuda_types.cuda can't break zluda_server_common.
- Position
- Below the diagonal: cuda_types.cuda uses a module listed above it, which is the healthy direction.
30→1 ptx.pass uses kernel_metadata✕
- Connections
- 2 pairs of types, where a type in ptx.pass uses a type in kernel_metadata.
- The types involved
ptx.pass.Module→kernel_metadata.ModuleMetadata32Bitptx.pass.Module→kernel_metadata.ModuleMetadataV1
- Direction
- ptx.pass uses kernel_metadata, so a change to kernel_metadata can break ptx.pass, while a change to ptx.pass can't break kernel_metadata.
- Position
- Below the diagonal: ptx.pass uses a module listed above it, which is the healthy direction.
30→2 ptx.pass uses llvm_zluda.utils✕
- Connections
- 3 pairs of types, where a type in ptx.pass uses a type in llvm_zluda.utils.
- The types involved
ptx.pass.Module→llvm_zluda.utils.Contextptx.pass.Module→llvm_zluda.utils.Moduleptx.pass.pass$module→llvm_zluda.utils.Module
- Direction
- ptx.pass uses llvm_zluda.utils, so a change to llvm_zluda.utils can break ptx.pass, while a change to ptx.pass can't break llvm_zluda.utils.
- Position
- Below the diagonal: ptx.pass uses a module listed above it, which is the healthy direction.
30→7 ptx.pass uses ptx_parser.ast✕
- Connections
- 8 pairs of types, where a type in ptx.pass uses a type in ptx_parser.ast.
- The types involved
ptx.pass.ConstantDefinition→ptx_parser.ast.ImmediateValueptx.pass.Function→ptx_parser.ast.LinkingDirectiveptx.pass.Function→ptx_parser.ast.TuningDirectiveptx.pass.Function→ptx_parser.ast.Variableptx.pass.ImplicitConversion→ptx_parser.ast.Typeptx.pass.PtrAccess→ptx_parser.ast.Typeptx.pass.PtxSpecialRegister→ptx_parser.ast.Typeptx.pass.SpirvWord→ptx_parser.ast.Operand
- Direction
- ptx.pass uses ptx_parser.ast, so a change to ptx_parser.ast can break ptx.pass, while a change to ptx.pass can't break ptx_parser.ast.
- Position
- Below the diagonal: ptx.pass uses a module listed above it, which is the healthy direction.
30→21 ptx.pass uses ptx_parser✕
- Connections
- 8 pairs of types, where a type in ptx.pass uses a type in ptx_parser.
- The types involved
ptx.pass.ConstantDefinition→ptx_parser.ScalarTypeptx.pass.ImplicitConversion→ptx_parser.StateSpaceptx.pass.PtrAccess→ptx_parser.StateSpaceptx.pass.PtxSpecialRegister→ptx_parser.ScalarTypeptx.pass.RepackVectorDetails→ptx_parser.ScalarTypeptx.pass.VectorRead→ptx_parser.ScalarTypeptx.pass.VectorWrite→ptx_parser.ScalarTypeptx.pass.pass$module→ptx_parser.ScalarType
- Direction
- ptx.pass uses ptx_parser, so a change to ptx_parser can break ptx.pass, while a change to ptx.pass can't break ptx_parser.
- Position
- Below the diagonal: ptx.pass uses a module listed above it, which is the healthy direction.
30→24 ptx.pass uses ptx.pass.instruction_mode_to_global_mode✕
- Connections
- 1 pair of types, where a type in ptx.pass uses a type in ptx.pass.instruction_mode_to_global_mode.
- The types involved
ptx.pass.KernelAttributes→ptx.pass.instruction_mode_to_global_mode.RoundingMode
- Direction
- ptx.pass uses ptx.pass.instruction_mode_to_global_mode, so a change to ptx.pass.instruction_mode_to_global_mode can break ptx.pass, while a change to ptx.pass can't break ptx.pass.instruction_mode_to_global_mode.
- Position
- Below the diagonal: ptx.pass uses a module listed above it, which is the healthy direction.
30→27 ptx.pass uses ptx.pass.llvm.emit✕
- Connections
- 1 pair of types, where a type in ptx.pass uses a type in ptx.pass.llvm.emit.
- The types involved
ptx.pass.pass$module→ptx.pass.llvm.emit.FloatingPointMode
- Direction
- ptx.pass uses ptx.pass.llvm.emit, so a change to ptx.pass.llvm.emit can break ptx.pass, while a change to ptx.pass can't break ptx.pass.llvm.emit.
- Position
- Below the diagonal: ptx.pass uses a module listed above it, which is the healthy direction.
31→7 ptx.pass.insert_explicit_load_store uses ptx_parser.ast✕
- Connections
- 6 pairs of types, where a type in ptx.pass.insert_explicit_load_store uses a type in ptx_parser.ast.
- The types involved
ptx.pass.insert_explicit_load_store.InsertMemSSAVisitor→ptx_parser.ast.LdDetailsptx.pass.insert_explicit_load_store.InsertMemSSAVisitor→ptx_parser.ast.MovDetailsptx.pass.insert_explicit_load_store.InsertMemSSAVisitor→ptx_parser.ast.StDataptx.pass.insert_explicit_load_store.InsertMemSSAVisitor→ptx_parser.ast.Typeptx.pass.insert_explicit_load_store.InsertMemSSAVisitor→ptx_parser.ast.Variableptx.pass.insert_explicit_load_store.InsertMemSSAVisitor→ptx_parser.ast.VisitorMap
- Direction
- ptx.pass.insert_explicit_load_store uses ptx_parser.ast, so a change to ptx_parser.ast can break ptx.pass.insert_explicit_load_store, while a change to ptx.pass.insert_explicit_load_store can't break ptx_parser.ast.
- Position
- Below the diagonal: ptx.pass.insert_explicit_load_store uses a module listed above it, which is the healthy direction.
31→18 ptx.pass.insert_explicit_load_store uses ptx_parser_macros_impl.parser✕
- Connections
- 1 pair of types, where a type in ptx.pass.insert_explicit_load_store uses a type in ptx_parser_macros_impl.parser.
- The types involved
ptx.pass.insert_explicit_load_store.InsertMemSSAVisitor→ptx_parser_macros_impl.parser.Instruction
- Direction
- ptx.pass.insert_explicit_load_store uses ptx_parser_macros_impl.parser, so a change to ptx_parser_macros_impl.parser can break ptx.pass.insert_explicit_load_store, while a change to ptx.pass.insert_explicit_load_store can't break ptx_parser_macros_impl.parser.
- Position
- Below the diagonal: ptx.pass.insert_explicit_load_store uses a module listed above it, which is the healthy direction.
31→21 ptx.pass.insert_explicit_load_store uses ptx_parser✕
- Connections
- 1 pair of types, where a type in ptx.pass.insert_explicit_load_store uses a type in ptx_parser.
- The types involved
ptx.pass.insert_explicit_load_store.InsertMemSSAVisitor→ptx_parser.StateSpace
- Direction
- ptx.pass.insert_explicit_load_store uses ptx_parser, so a change to ptx_parser can break ptx.pass.insert_explicit_load_store, while a change to ptx.pass.insert_explicit_load_store can't break ptx_parser.
- Position
- Below the diagonal: ptx.pass.insert_explicit_load_store uses a module listed above it, which is the healthy direction.
31→30 ptx.pass.insert_explicit_load_store uses ptx.pass✕
- Connections
- 6 pairs of types, where a type in ptx.pass.insert_explicit_load_store uses a type in ptx.pass.
- The types involved
ptx.pass.insert_explicit_load_store.InsertMemSSAVisitor→ptx.pass.GlobalStringIdentResolver2ptx.pass.insert_explicit_load_store.InsertMemSSAVisitor→ptx.pass.PtrAccessptx.pass.insert_explicit_load_store.InsertMemSSAVisitor→ptx.pass.SpirvWordptx.pass.insert_explicit_load_store.insert_explicit_load_store$module→ptx.pass.Directive2ptx.pass.insert_explicit_load_store.insert_explicit_load_store$module→ptx.pass.Functionptx.pass.insert_explicit_load_store.insert_explicit_load_store$module→ptx.pass.GlobalStringIdentResolver2
- Direction
- ptx.pass.insert_explicit_load_store uses ptx.pass, so a change to ptx.pass can break ptx.pass.insert_explicit_load_store, while a change to ptx.pass.insert_explicit_load_store can't break ptx.pass.
- Position
- Below the diagonal: ptx.pass.insert_explicit_load_store uses a module listed above it, which is the healthy direction.
32→2 zluda.impl.context uses llvm_zluda.utils✕
- Connections
- 1 pair of types, where a type in zluda.impl.context uses a type in llvm_zluda.utils.
- The types involved
zluda.impl.context.context$module→llvm_zluda.utils.Context
- Direction
- zluda.impl.context uses llvm_zluda.utils, so a change to llvm_zluda.utils can break zluda.impl.context, while a change to zluda.impl.context can't break llvm_zluda.utils.
- Position
- Below the diagonal: zluda.impl.context uses a module listed above it, which is the healthy direction.
32→3 zluda.impl.context uses zluda_common✕
- Connections
- 2 pairs of types, where a type in zluda.impl.context uses a type in zluda_common.
- The types involved
zluda.impl.context.Context→zluda_common.ZludaObjectzluda.impl.context.MaybeContextRef→zluda_common.FromCuda
- Direction
- zluda.impl.context uses zluda_common, so a change to zluda_common can break zluda.impl.context, while a change to zluda.impl.context can't break zluda_common.
- Position
- Below the diagonal: zluda.impl.context uses a module listed above it, which is the healthy direction.
32→6 zluda.impl.context uses hip_runtime_sys✕
- Connections
- 2 pairs of types, where a type in zluda.impl.context uses a type in hip_runtime_sys.
- The types involved
zluda.impl.context.Context→hip_runtime_sys.hipStream_tzluda.impl.context.context$module→hip_runtime_sys.hipLimit_t
- Direction
- zluda.impl.context uses hip_runtime_sys, so a change to hip_runtime_sys can break zluda.impl.context, while a change to zluda.impl.context can't break hip_runtime_sys.
- Position
- Below the diagonal: zluda.impl.context uses a module listed above it, which is the healthy direction.
32→29 zluda.impl.context uses cuda_types.cuda✕
- Connections
- 2 pairs of types, where a type in zluda.impl.context uses a type in cuda_types.cuda.
- The types involved
zluda.impl.context.StorageData→cuda_types.cuda.CUcontextzluda.impl.context.context$module→cuda_types.cuda.CUcontext
- Direction
- zluda.impl.context uses cuda_types.cuda, so a change to cuda_types.cuda can break zluda.impl.context, while a change to zluda.impl.context can't break cuda_types.cuda.
- Position
- Below the diagonal: zluda.impl.context uses a module listed above it, which is the healthy direction.
33→3 zluda.impl.function uses zluda_common✕
- Connections
- 1 pair of types, where a type in zluda.impl.function uses a type in zluda_common.
- The types involved
zluda.impl.function.Function→zluda_common.FromCuda
- Direction
- zluda.impl.function uses zluda_common, so a change to zluda_common can break zluda.impl.function, while a change to zluda.impl.function can't break zluda_common.
- Position
- Below the diagonal: zluda.impl.function uses a module listed above it, which is the healthy direction.
33→6 zluda.impl.function uses hip_runtime_sys✕
- Connections
- 2 pairs of types, where a type in zluda.impl.function uses a type in hip_runtime_sys.
- The types involved
zluda.impl.function.Function→hip_runtime_sys.hipFunction_tzluda.impl.function.function$module→hip_runtime_sys.hipStream_t
- Direction
- zluda.impl.function uses hip_runtime_sys, so a change to hip_runtime_sys can break zluda.impl.function, while a change to zluda.impl.function can't break hip_runtime_sys.
- Position
- Below the diagonal: zluda.impl.function uses a module listed above it, which is the healthy direction.
33→12 zluda.impl.function uses dark_api✕
- Connections
- 1 pair of types, where a type in zluda.impl.function uses a type in dark_api.
- The types involved
zluda.impl.function.Function→dark_api.FunctionArgInfo
- Direction
- zluda.impl.function uses dark_api, so a change to dark_api can break zluda.impl.function, while a change to zluda.impl.function can't break dark_api.
- Position
- Below the diagonal: zluda.impl.function uses a module listed above it, which is the healthy direction.
33→29 zluda.impl.function uses cuda_types.cuda✕
- Connections
- 1 pair of types, where a type in zluda.impl.function uses a type in cuda_types.cuda.
- The types involved
zluda.impl.function.Function→cuda_types.cuda.CUfunction
- Direction
- zluda.impl.function uses cuda_types.cuda, so a change to cuda_types.cuda can break zluda.impl.function, while a change to zluda.impl.function can't break cuda_types.cuda.
- Position
- Below the diagonal: zluda.impl.function uses a module listed above it, which is the healthy direction.
33→30 zluda.impl.function uses ptx.pass✕
- Connections
- 1 pair of types, where a type in zluda.impl.function uses a type in ptx.pass.
- The types involved
zluda.impl.function.function$module→ptx.pass.Function
- Direction
- zluda.impl.function uses ptx.pass, so a change to ptx.pass can break zluda.impl.function, while a change to zluda.impl.function can't break ptx.pass.
- Position
- Below the diagonal: zluda.impl.function uses a module listed above it, which is the healthy direction.
34→3 zluda.impl.library uses zluda_common✕
- Connections
- 1 pair of types, where a type in zluda.impl.library uses a type in zluda_common.
- The types involved
zluda.impl.library.Library→zluda_common.ZludaObject
- Direction
- zluda.impl.library uses zluda_common, so a change to zluda_common can break zluda.impl.library, while a change to zluda.impl.library can't break zluda_common.
- Position
- Below the diagonal: zluda.impl.library uses a module listed above it, which is the healthy direction.
34→29 zluda.impl.library uses cuda_types.cuda✕
- Connections
- 2 pairs of types, where a type in zluda.impl.library uses a type in cuda_types.cuda.
- The types involved
zluda.impl.library.library$module→cuda_types.cuda.CUlibraryzluda.impl.library.library$module→cuda_types.cuda.CUmodule
- Direction
- zluda.impl.library uses cuda_types.cuda, so a change to cuda_types.cuda can break zluda.impl.library, while a change to zluda.impl.library can't break cuda_types.cuda.
- Position
- Below the diagonal: zluda.impl.library uses a module listed above it, which is the healthy direction.
34→30 zluda.impl.library uses ptx.pass✕
- Connections
- 1 pair of types, where a type in zluda.impl.library uses a type in ptx.pass.
- The types involved
zluda.impl.library.library$module→ptx.pass.Function
- Direction
- zluda.impl.library uses ptx.pass, so a change to ptx.pass can break zluda.impl.library, while a change to zluda.impl.library can't break ptx.pass.
- Position
- Below the diagonal: zluda.impl.library uses a module listed above it, which is the healthy direction.
35→34 compiler.hip uses zluda.impl.library✕
- Connections
- 1 pair of types, where a type in compiler.hip uses a type in zluda.impl.library.
- The types involved
compiler.hip.Runtime→zluda.impl.library.Library
- Direction
- compiler.hip uses zluda.impl.library, so a change to zluda.impl.library can break compiler.hip, while a change to compiler.hip can't break zluda.impl.library.
- Position
- Below the diagonal: compiler.hip uses a module listed above it, which is the healthy direction.
36→2 zluda.impl.graphics uses llvm_zluda.utils✕
- Connections
- 1 pair of types, where a type in zluda.impl.graphics uses a type in llvm_zluda.utils.
- The types involved
zluda.impl.graphics.OpenCLContext→llvm_zluda.utils.Context
- Direction
- zluda.impl.graphics uses llvm_zluda.utils, so a change to llvm_zluda.utils can break zluda.impl.graphics, while a change to zluda.impl.graphics can't break llvm_zluda.utils.
- Position
- Below the diagonal: zluda.impl.graphics uses a module listed above it, which is the healthy direction.
36→6 zluda.impl.graphics uses hip_runtime_sys✕
- Connections
- 1 pair of types, where a type in zluda.impl.graphics uses a type in hip_runtime_sys.
- The types involved
zluda.impl.graphics.graphics$module→hip_runtime_sys.hipStream_t
- Direction
- zluda.impl.graphics uses hip_runtime_sys, so a change to hip_runtime_sys can break zluda.impl.graphics, while a change to zluda.impl.graphics can't break hip_runtime_sys.
- Position
- Below the diagonal: zluda.impl.graphics uses a module listed above it, which is the healthy direction.
36→29 zluda.impl.graphics uses cuda_types.cuda✕
- Connections
- 2 pairs of types, where a type in zluda.impl.graphics uses a type in cuda_types.cuda.
- The types involved
zluda.impl.graphics.graphics$module→cuda_types.cuda.ID3D11Devicezluda.impl.graphics.graphics$module→cuda_types.cuda.ID3D11Resource
- Direction
- zluda.impl.graphics uses cuda_types.cuda, so a change to cuda_types.cuda can break zluda.impl.graphics, while a change to zluda.impl.graphics can't break cuda_types.cuda.
- Position
- Below the diagonal: zluda.impl.graphics uses a module listed above it, which is the healthy direction.
36→34 zluda.impl.graphics uses zluda.impl.library✕
- Connections
- 1 pair of types, where a type in zluda.impl.graphics uses a type in zluda.impl.library.
- The types involved
zluda.impl.graphics.OpenCLContext→zluda.impl.library.Library
- Direction
- zluda.impl.graphics uses zluda.impl.library, so a change to zluda.impl.library can break zluda.impl.graphics, while a change to zluda.impl.graphics can't break zluda.impl.library.
- Position
- Below the diagonal: zluda.impl.graphics uses a module listed above it, which is the healthy direction.
37→3 zluda_blas.impl uses zluda_common✕
- Connections
- 2 pairs of types, where a type in zluda_blas.impl uses a type in zluda_common.
- The types involved
zluda_blas.impl.Handle→zluda_common.FromCudazluda_blas.impl.Handle→zluda_common.ZludaObject
- Direction
- zluda_blas.impl uses zluda_common, so a change to zluda_common can break zluda_blas.impl, while a change to zluda_blas.impl can't break zluda_common.
- Position
- Below the diagonal: zluda_blas.impl uses a module listed above it, which is the healthy direction.
37→6 zluda_blas.impl uses hip_runtime_sys✕
- Connections
- 1 pair of types, where a type in zluda_blas.impl uses a type in hip_runtime_sys.
- The types involved
zluda_blas.impl.impl$module→hip_runtime_sys.hipStream_t
- Direction
- zluda_blas.impl uses hip_runtime_sys, so a change to hip_runtime_sys can break zluda_blas.impl, while a change to zluda_blas.impl can't break hip_runtime_sys.
- Position
- Below the diagonal: zluda_blas.impl uses a module listed above it, which is the healthy direction.
37→19 zluda_blas.impl uses cuda_types.cublas✕
- Connections
- 2 pairs of types, where a type in zluda_blas.impl uses a type in cuda_types.cublas.
- The types involved
zluda_blas.impl.impl$module→cuda_types.cublas.cublasHandle_tzluda_blas.impl.impl$module→cuda_types.cublas.cublasMath_t
- Direction
- zluda_blas.impl uses cuda_types.cublas, so a change to cuda_types.cublas can break zluda_blas.impl, while a change to zluda_blas.impl can't break cuda_types.cublas.
- Position
- Below the diagonal: zluda_blas.impl uses a module listed above it, which is the healthy direction.
37→34 zluda_blas.impl uses zluda.impl.library✕
- Connections
- 1 pair of types, where a type in zluda_blas.impl uses a type in zluda.impl.library.
- The types involved
zluda_blas.impl.Handle→zluda.impl.library.Library
- Direction
- zluda_blas.impl uses zluda.impl.library, so a change to zluda.impl.library can break zluda_blas.impl, while a change to zluda_blas.impl can't break zluda.impl.library.
- Position
- Below the diagonal: zluda_blas.impl uses a module listed above it, which is the healthy direction.
38→3 zluda.impl.driver uses zluda_common✕
- Connections
- 1 pair of types, where a type in zluda.impl.driver uses a type in zluda_common.
- The types involved
zluda.impl.driver.Device→zluda_common.LiveCheck
- Direction
- zluda.impl.driver uses zluda_common, so a change to zluda_common can break zluda.impl.driver, while a change to zluda.impl.driver can't break zluda_common.
- Position
- Below the diagonal: zluda.impl.driver uses a module listed above it, which is the healthy direction.
38→16 zluda.impl.driver uses dark_api.cuda✕
- Connections
- 2 pairs of types, where a type in zluda.impl.driver uses a type in dark_api.cuda.
- The types involved
zluda.impl.driver.DarkApi→dark_api.cuda.CudaDarkApizluda.impl.driver.Zluda32DarkApi→dark_api.cuda.CudaDarkApi
- Direction
- zluda.impl.driver uses dark_api.cuda, so a change to dark_api.cuda can break zluda.impl.driver, while a change to zluda.impl.driver can't break dark_api.cuda.
- Position
- Below the diagonal: zluda.impl.driver uses a module listed above it, which is the healthy direction.
38→29 zluda.impl.driver uses cuda_types.cuda✕
- Connections
- 5 pairs of types, where a type in zluda.impl.driver uses a type in cuda_types.cuda.
- The types involved
zluda.impl.driver.AllocationInfo→cuda_types.cuda.CUcontextzluda.impl.driver.Allocations→cuda_types.cuda.CUcontextzluda.impl.driver.DarkApi→cuda_types.cuda.CUcontextzluda.impl.driver.Zluda32DarkApi→cuda_types.cuda.CUfunctionzluda.impl.driver.Zluda32DarkApi→cuda_types.cuda.CUmodule
- Direction
- zluda.impl.driver uses cuda_types.cuda, so a change to cuda_types.cuda can break zluda.impl.driver, while a change to zluda.impl.driver can't break cuda_types.cuda.
- Position
- Below the diagonal: zluda.impl.driver uses a module listed above it, which is the healthy direction.
38→30 zluda.impl.driver uses ptx.pass✕
- Connections
- 1 pair of types, where a type in zluda.impl.driver uses a type in ptx.pass.
- The types involved
zluda.impl.driver.driver$module→ptx.pass.Function
- Direction
- zluda.impl.driver uses ptx.pass, so a change to ptx.pass can break zluda.impl.driver, while a change to zluda.impl.driver can't break ptx.pass.
- Position
- Below the diagonal: zluda.impl.driver uses a module listed above it, which is the healthy direction.
38→36 zluda.impl.driver uses zluda.impl.graphics✕
- Connections
- 1 pair of types, where a type in zluda.impl.driver uses a type in zluda.impl.graphics.
- The types involved
zluda.impl.driver.Device→zluda.impl.graphics.OpenCLContext
- Direction
- zluda.impl.driver uses zluda.impl.graphics, so a change to zluda.impl.graphics can break zluda.impl.driver, while a change to zluda.impl.driver can't break zluda.impl.graphics.
- Position
- Below the diagonal: zluda.impl.driver uses a module listed above it, which is the healthy direction.
39→1 zluda.impl.module uses kernel_metadata✕
- Connections
- 2 pairs of types, where a type in zluda.impl.module uses a type in kernel_metadata.
- The types involved
zluda.impl.module.Metadata32Bit→kernel_metadata.Global32Bitzluda.impl.module.Metadata32Bit→kernel_metadata.ModuleMetadata32Bit
- Direction
- zluda.impl.module uses kernel_metadata, so a change to kernel_metadata can break zluda.impl.module, while a change to zluda.impl.module can't break kernel_metadata.
- Position
- Below the diagonal: zluda.impl.module uses a module listed above it, which is the healthy direction.
39→2 zluda.impl.module uses llvm_zluda.utils✕
- Connections
- 1 pair of types, where a type in zluda.impl.module uses a type in llvm_zluda.utils.
- The types involved
zluda.impl.module.module$module→llvm_zluda.utils.Module
- Direction
- zluda.impl.module uses llvm_zluda.utils, so a change to llvm_zluda.utils can break zluda.impl.module, while a change to zluda.impl.module can't break llvm_zluda.utils.
- Position
- Below the diagonal: zluda.impl.module uses a module listed above it, which is the healthy direction.
39→3 zluda.impl.module uses zluda_common✕
- Connections
- 2 pairs of types, where a type in zluda.impl.module uses a type in zluda_common.
- The types involved
zluda.impl.module.Module→zluda_common.ZludaObjectzluda.impl.module.module$module→zluda_common.CodeLibraryRef
- Direction
- zluda.impl.module uses zluda_common, so a change to zluda_common can break zluda.impl.module, while a change to zluda.impl.module can't break zluda_common.
- Position
- Below the diagonal: zluda.impl.module uses a module listed above it, which is the healthy direction.
39→6 zluda.impl.module uses hip_runtime_sys✕
- Connections
- 2 pairs of types, where a type in zluda.impl.module uses a type in hip_runtime_sys.
- The types involved
zluda.impl.module.Module→hip_runtime_sys.hipModule_tzluda.impl.module.ModuleMutable→hip_runtime_sys.hipModule_t
- Direction
- zluda.impl.module uses hip_runtime_sys, so a change to hip_runtime_sys can break zluda.impl.module, while a change to zluda.impl.module can't break hip_runtime_sys.
- Position
- Below the diagonal: zluda.impl.module uses a module listed above it, which is the healthy direction.
39→12 zluda.impl.module uses dark_api✕
- Connections
- 1 pair of types, where a type in zluda.impl.module uses a type in dark_api.
- The types involved
zluda.impl.module.ModuleMutable→dark_api.FunctionArgInfo
- Direction
- zluda.impl.module uses dark_api, so a change to dark_api can break zluda.impl.module, while a change to zluda.impl.module can't break dark_api.
- Position
- Below the diagonal: zluda.impl.module uses a module listed above it, which is the healthy direction.
39→29 zluda.impl.module uses cuda_types.cuda✕
- Connections
- 2 pairs of types, where a type in zluda.impl.module uses a type in cuda_types.cuda.
- The types involved
zluda.impl.module.module$module→cuda_types.cuda.CUjit_option_enumzluda.impl.module.module$module→cuda_types.cuda.CUmodule
- Direction
- zluda.impl.module uses cuda_types.cuda, so a change to cuda_types.cuda can break zluda.impl.module, while a change to zluda.impl.module can't break cuda_types.cuda.
- Position
- Below the diagonal: zluda.impl.module uses a module listed above it, which is the healthy direction.
39→30 zluda.impl.module uses ptx.pass✕
- Connections
- 1 pair of types, where a type in zluda.impl.module uses a type in ptx.pass.
- The types involved
zluda.impl.module.module$module→ptx.pass.Function
- Direction
- zluda.impl.module uses ptx.pass, so a change to ptx.pass can break zluda.impl.module, while a change to zluda.impl.module can't break ptx.pass.
- Position
- Below the diagonal: zluda.impl.module uses a module listed above it, which is the healthy direction.
39→35 zluda.impl.module uses compiler.hip✕
- Connections
- 1 pair of types, where a type in zluda.impl.module uses a type in compiler.hip.
- The types involved
zluda.impl.module.module$module→compiler.hip.hipDeviceProp_tR0600
- Direction
- zluda.impl.module uses compiler.hip, so a change to compiler.hip can break zluda.impl.module, while a change to zluda.impl.module can't break compiler.hip.
- Position
- Below the diagonal: zluda.impl.module uses a module listed above it, which is the healthy direction.
40→3 zluda_blaslt.impl uses zluda_common✕
- Connections
- 3 pairs of types, where a type in zluda_blaslt.impl uses a type in zluda_common.
- The types involved
zluda_blaslt.impl.BlasHandle→zluda_common.ZludaObjectzluda_blaslt.impl.Handle→zluda_common.FromCudazluda_blaslt.impl.Handle→zluda_common.ZludaObject
- Direction
- zluda_blaslt.impl uses zluda_common, so a change to zluda_common can break zluda_blaslt.impl, while a change to zluda_blaslt.impl can't break zluda_common.
- Position
- Below the diagonal: zluda_blaslt.impl uses a module listed above it, which is the healthy direction.
40→6 zluda_blaslt.impl uses hip_runtime_sys✕
- Connections
- 1 pair of types, where a type in zluda_blaslt.impl uses a type in hip_runtime_sys.
- The types involved
zluda_blaslt.impl.impl$module→hip_runtime_sys.hipStream_t
- Direction
- zluda_blaslt.impl uses hip_runtime_sys, so a change to hip_runtime_sys can break zluda_blaslt.impl, while a change to zluda_blaslt.impl can't break hip_runtime_sys.
- Position
- Below the diagonal: zluda_blaslt.impl uses a module listed above it, which is the healthy direction.
40→8 zluda_blaslt.impl uses cuda_types.cublaslt✕
- Connections
- 4 pairs of types, where a type in zluda_blaslt.impl uses a type in cuda_types.cublaslt.
- The types involved
zluda_blaslt.impl.impl$module→cuda_types.cublaslt.cublasLtMatmulAlgo_tzluda_blaslt.impl.impl$module→cuda_types.cublaslt.cublasLtMatmulDescAttributes_tzluda_blaslt.impl.impl$module→cuda_types.cublaslt.cublasLtMatmulHeuristicResult_tzluda_blaslt.impl.impl$module→cuda_types.cublaslt.cublasLtMatmulPreferenceAttributes_t
- Direction
- zluda_blaslt.impl uses cuda_types.cublaslt, so a change to cuda_types.cublaslt can break zluda_blaslt.impl, while a change to zluda_blaslt.impl can't break cuda_types.cublaslt.
- Position
- Below the diagonal: zluda_blaslt.impl uses a module listed above it, which is the healthy direction.
40→17 zluda_blaslt.impl uses hipblaslt_sys✕
- Connections
- 8 pairs of types, where a type in zluda_blaslt.impl uses a type in hipblaslt_sys.
- The types involved
zluda_blaslt.impl.Handle→hipblaslt_sys.hipblasLtHandle_tzluda_blaslt.impl.impl$module→hipblaslt_sys.hipDataTypezluda_blaslt.impl.impl$module→hipblaslt_sys.hipblasComputeType_tzluda_blaslt.impl.impl$module→hipblaslt_sys.hipblasLtMatmulDescAttributes_tzluda_blaslt.impl.impl$module→hipblaslt_sys.hipblasLtMatmulDesc_tzluda_blaslt.impl.impl$module→hipblaslt_sys.hipblasLtMatmulPreference_tzluda_blaslt.impl.impl$module→hipblaslt_sys.hipblasLtMatrixLayoutAttribute_tzluda_blaslt.impl.impl$module→hipblaslt_sys.hipblasLtMatrixLayout_t
- Direction
- zluda_blaslt.impl uses hipblaslt_sys, so a change to hipblaslt_sys can break zluda_blaslt.impl, while a change to zluda_blaslt.impl can't break hipblaslt_sys.
- Position
- Below the diagonal: zluda_blaslt.impl uses a module listed above it, which is the healthy direction.
40→37 zluda_blaslt.impl uses zluda_blas.impl✕
- Connections
- 1 pair of types, where a type in zluda_blaslt.impl uses a type in zluda_blas.impl.
- The types involved
zluda_blaslt.impl.impl$module→zluda_blas.impl.Handle
- Direction
- zluda_blaslt.impl uses zluda_blas.impl, so a change to zluda_blas.impl can break zluda_blaslt.impl, while a change to zluda_blaslt.impl can't break zluda_blas.impl.
- Position
- Below the diagonal: zluda_blaslt.impl uses a module listed above it, which is the healthy direction.
At a glance — Code Health
At a glance — Architecture
At a glance — Maturity
At a glance — Readiness
At a glance — Security
At a glance — Domain Modelling
Security & Compliance — OWASP Top-10 mapping
| OWASP category | Findings | Severity |
|---|---|---|
| A05:2025 — Injection | 45 | High / Critical |
| A03:2025 — Software Supply Chain Failures | 36 | High / Critical |
| A02:2025 — Security Misconfiguration | 4 | Medium |
Roadmap
First, enable nullable reference types across all projects and resolve warnings directly instead of suppressing them. Next, address the single missing architecture decision record and begin refactoring the twenty-eight most complex code bodies to reduce cognitive load. Finally, document significant design decisions in a dedicated folder and expand the main README with getting-started instructions and an architecture overview.
| Do this | Helps | Effort | Dimension |
|---|---|---|---|
| Resolve the 1 No ADRs found finding(s) in ADR Quality. | +4.7 pts | Low | ADR Quality |
| Enable <Nullable>enable</Nullable> across all projects and resolve warnings rather than suppressing with `!`. | +7.7 pts | Medium | Nullable reference types |
| Bring the 28 bodies over 15 down to 15 or less in Cognitive Complexity — start with zluda_bindgen::cuda_derive_display_trait_for_item (cognitive 62), OpcodeDefinitions::new (cognitive 58), dark_api::cuda::guid_to_name (cognitive 46). | +4.0 pts | High | Cognitive Complexity |
| Resolve the 1 Orphaned knowledge finding(s) in Knowledge Freshness — start with lib.rs. | +1.5 pts | Low | Knowledge Freshness |
| Record your significant design decisions. Write one short document per decision, giving its date, the situation, the decision and its consequences, and keep them together with your other design documents. A docs/adr folder with each file named by number and title is the easiest place to find them. | +2.7 pts | Medium | Architecture documentation |
| Expand the README with getting-started, architecture overview and a project map. | +2.6 pts | Medium | Documentation (README) |
| Resolve the 1 Off-boarding risk finding(s) in Bus Factor. | +0.9 pts | Low | Bus Factor |
| Bring the 1 body over 30 down to 30 or less in Cyclomatic Complexity — start with ptx::pass::replace_instructions_with_functions::run_instruction (cyclomatic 60). | +0.8 pts | Medium | Cyclomatic Complexity |
File quality
| File | Score | Band | Worst signal |
|---|---|---|---|
| REDACTED | 2.3 | Slop | Static Analysis (SAST): High: REDACTED |
| REDACTED | 2.3 | Slop | Static Analysis (SAST): High: REDACTED |
| REDACTED | 2.4 | Slop | Static Analysis (SAST): High: REDACTED |
| REDACTED | 3.0 | Slop | Static Analysis (SAST): High: REDACTED |
| REDACTED | 3.0 | Slop | Dependency Vulnerabilities: High vulnerability: REDACTED |
| zluda_bindgen/src/main.rs | 7.0 | Mixed | Cyclomatic Complexity: zluda_bindgen::cuda_derive_display_trait_for_item (cyclomatic 30) |
| ptx/src/pass/llvm/emit.rs | 7.0 | Mixed | Cyclomatic Complexity: ModuleEmitContext::emit_method (cyclomatic 18) |
| ptx/src/pass/instruction_mode_to_global_mode/mod.rs | 7.0 | Mixed | Cognitive Complexity: PartialModeInsertion::new (cognitive 28) |
| ptx_parser_macros/src/lib.rs | 7.0 | Mixed | Cyclomatic Complexity: ptx_parser_macros::emit_definition_parser (cyclomatic 23) |
| zluda/src/impl/driver.rs | 7.0 | Mixed | Cognitive Complexity: Zluda32DarkApi::get_module_globals (cognitive 19) |
| format/src/lib.rs | 7.0 | Mixed | God Classes: FileTooLong: src/lib.rs |
| format/src/dnn9.rs | 7.0 | Mixed | God Classes: FunctionTooLong: format::dnn9::cudnn9_print_elements |
| ptx/src/pass/insert_implicit_conversions.rs | 7.1 | Mixed | Cyclomatic Complexity: ptx::pass::insert_implicit_conversions::should_convert_relaxed_dst (cyclomatic 17) |
| ptx/src/pass/deparamize_functions.rs | 7.1 | Mixed | Cognitive Complexity: ptx::pass::deparamize_functions::run_method (cognitive 20) |
| nvapi/src/nvapi.rs | 7.1 | Mixed | Code Duplication: Members sharing a duplicated core (59 members, 50+ identical tokens) |
| ptx/src/pass/replace_instructions_with_functions.rs | 7.1 | Mixed | Cyclomatic Complexity: ptx::pass::replace_instructions_with_functions::run_instruction (cyclomatic 60) |
| ptx_parser/src/lib.rs | 7.2 | Mixed | Cognitive Complexity: ptx_parser::multi_variable (cognitive 17) |
| zluda_trace/src/lib.rs | 7.2 | Mixed | God Classes: FileTooLong: src/lib.rs |
| REDACTED | 7.2 | Mixed | Static Analysis (SAST): High: REDACTED |
| ptx_parser/src/ast.rs | 7.4 | Mixed | Cyclomatic Complexity: CvtDetails::new (cyclomatic 23) |
How the grades work
A definite problem that already costs you something and drags the score down: a missing authorisation check, a dependency with a known exploit, a build that does not reproduce. Failure here tends to cause failures elsewhere.
Likely wrong, but not failing yet. It degrades the codebase over a longer horizon and can cause failures elsewhere — not urgent this week, not something to carry for two years either.
Recorded, with no effect on how the codebase functions. Present so the survey is complete, not because it needs doing.
Something this survey could not settle from the outside, and which could be critical or serious. Either a control was required and no positive evidence of it exists in the repository — a backup job that nothing shows was ever restored from proves nothing about restores — or our own analysis could not run over that part of the tree. This is not a clean result. These are excluded from the score rather than awarded a pass, so the number on the cover neither rewards nor penalises them: if you act on this survey without resolving them, you carry that risk yourself. Each one is named under Limitations.
Methodology & how to trust this report
Watchdog is a deep, periodic assessment — run each sprint, monthly, or quarterly, taking the time to go wider and deeper than a quick check and surfacing in one coherent report what you'd otherwise piece together from a dozen separate tools. It scores deterministically: the same commit yields the same score, every run. 59 of 59 evaluated dimensions are computed purely by tools and static analysis (confidence 1.0); none rely on LLM judgement. Overall confidence is 0.8 — the weighted average across measured dimensions; it falls as more of the score leans on LLM-assisted judgement and rises when it's fully tool-backed.
What we checked — 59 dimensions across the health lenses
- Can you open the finding? Real findings cite a repo-relative file and line you can open at the cited line — never an absolute scratch path. Here, 237 of 263 do; the remainder are repo-wide signals — a dimension-level measurement, not a single line.
- Is there a tool behind the number? Every score below names the method that produced it — Roslyn, git, a scanner, or (for a handful of documentation/naming dimensions) an LLM labelled sampled · advisory — not a narrative.
- Does re-running give the same result? Run it again on the same commit and the score — and this report, byte for byte — is identical. A report whose numbers move between runs is describing the run, not the code.
Tools & methods
| Method | Backs | Version | Evaluator |
|---|---|---|---|
| Roslyn static analysis | Complexity, cohesion, coupling, dead code, API surface, layering | 5.3.0 | ✓ deterministic |
| Native secret scanner | Hardcoded secrets / credentials | 1.0.0 | ✓ deterministic |
| Watchdog duplication detector (in-process) | Code duplication | 1.0.0 | ✓ deterministic |
| Coverage (coverlet / dotnet-coverage) | Line & branch coverage | 10.0.400 | ✓ deterministic |
| NuGet / dotnet | Outdated, vulnerable & deprecated dependencies | 10.0.400 | ✓ deterministic |
| git / LibGit2Sharp | Churn hotspots, knowledge concentration, history | 2.43.0 · 0.31.0 | ✓ deterministic |
| gitleaks · semgrep · trivy · checkov | Secrets in history, SAST, CVEs, IaC & container, PII / GDPR | 1.86.0 · 0.69.3 · 3.2.533 | ✓ deterministic |
| LLM (sampled · advisory) | Documentation quality, ADR conformance, naming — sampled over a bounded sample; advisory, never a deterministic measurement | — | ◐ LLM · sampled · advisory |
Run transparency — what happened this run
- D8 Code Coverage — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. Coverage NOT MEASURED: test source is present (.rs) but the built-in coverage collector has no runner for this repository's ecosystem — so this suite was never executed by it. Not scored — this is a gap in the analyzer's language coverage, not a defect in the repo. To have real coverage read, produce a coverage report in a standard format (lcov — `cargo install cargo-llvm-cov`, then `cargo llvm-cov --lcov --output-path lcov.info`) and commit it — a hosted scan measures a clone of the repository, so a report that exists only in a working tree, a CI runner's or your own, never reaches it; the artefact is commonly gitignored, so `git add -f` that one file (or un-ignore its path) and commit it alongside the code it measures. You can widen what we reach: optional: produce a coverage report in a standard format (lcov — `cargo install cargo-llvm-cov`, then `cargo llvm-cov --lcov --output-path lcov.info`) and commit it — a hosted scan measures a clone of the repository, so a report that exists only in a working tree, a CI runner's or your own, never reaches it; the artefact is commonly gitignored, so `git add -f` that one file (or un-ignore its path) and commit it alongside the code it measures — then the real number is read on the next scan.
- D11 Test Reliability — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. Test source is present (.rs) and this repository declares a Cargo test suite (repository root, 21 test files), but it was not re-run: no test result was produced. Not scored — this is a gap in the analyzer's language coverage, not a finding about this repository.
- D12 Dependency Hygiene — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. `dotnet list package` did not return usable output (restore or query failed); no outdated/vulnerable/deprecated data to assess. The tool said: `dotnet list package` exited 1 with no output.
- D19 Documentation Quality — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. Each ecosystem's manifests (what makes a folder runnable or a published package) need their own proof before D19 judges its units. Until then those folders' READMEs are judged only as documents (boilerplate, structure), never counted as complete.
- D19 Documentation Quality — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. This kind did not reach the 95 % precision gate in three adjudicated runs (docs/plans/deterministic-engine/adjudication/D19-sample3-v3.md), so by the pre-registered fallback it publishes no row until a re-measurement shows it does. The candidates are counted here, not reported as findings, and the score does not include them.
- D21 Naming Consistency — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. D21's deterministic naming rules are gated on C# first (docs/plans/deterministic-engine.md, P5): each other language is switched on only after its own concept suite and A/B gate pass. Until then its declarations carry no naming verdict this run — this says nothing about whether its names are consistent.
- D22 Internal API Consistency — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. D22's deterministic call-shape rules are gated on C# first (docs/plans/deterministic-engine.md, P5): each other language is switched on only after its own concept suite and A/B gate pass. This says nothing about whether that surface is consistent.
- D24 Comment Value — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. This check's reader does not cover the language this repository's product is written in, so it had nothing of the product to read. That is a gap in this analyzer's language reach — not a finding about this repository.
- D44 Platform End-of-Life — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. This dimension reads a project's own statement about the platform it runs on: a TargetFramework in a .NET project file, a .nvmrc or .python-version, a capped requires-python, a Rust toolchain file or Cargo.toml rust-version, a .go-version, .java-version, .ruby-version, .tool-versions or .sdkmanrc, a go.mod go directive, a Maven or Gradle Java level or toolchain, a Gemfile's ruby directive, a mix.exs elixir requirement, a rebar.config minimum_otp_vsn, a pubspec.yaml SDK constraint, a build.sbt scalaVersion, or a framework major pinned by a dependency constraint. This repository carries none of them, so nothing about its platform was established. That is a gap in this analyzer's coverage, NOT a finding that the platform is supported — a language whose runtime is declared elsewhere (Package.swift, a REDACTED) is simply not read here yet.
- AXR1 Runtime accessibility — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. Render strategy: boot via REDACTED (REDACTED). docker build failed (exit 1) — -------------------- ERROR: failed to solve: nvidia/cuda:13.0.1-base-ubuntu24.04: failed to resolve source metadata for docker.io/nvidia/cuda:13.0.1-base-ubuntu24.04: failed to do request: Head "https://registry-1.docker.io/v2/nvidia/cuda/manifests/13.0.1-base-ubuntu24.04": Forbidden; runtime evidence skipped.
- IC1 Incompleteness & stubs — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. This check is implemented over the C# syntax tree, and no C# was loaded on this run, so it had nothing to read. That is a gap in this analyzer's language reach — not a finding that the repository is free of what this check looks for.
- M4 Documentation accuracy — measured, with a gap in what it reached — Watchdog measured this, but not all of it. What it did not reach is a gap on our side — a collector, parser or image we have not built yet — so the numbers on that dimension cover less than the repository, and the part left out is not evidence that it would have passed. M4 checks only the claim kinds adjudicated at or above 95 % precision: links and versions (PREREGISTRATION version 9). Paths, command arguments, installs, environment variables, settings, projects, packages, routes, cd targets, infrastructure products and removed modules are read but not checked, and none of them is reported or counted.
- X1 Async correctness — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. This check is implemented over the C# syntax tree, and no C# was loaded on this run, so it had nothing to read. That is a gap in this analyzer's language reach — not a finding that the repository is free of what this check looks for.
- X2 Cancellation propagation — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. This check is implemented over the C# syntax tree, and no C# was loaded on this run, so it had nothing to read. That is a gap in this analyzer's language reach — not a finding that the repository is free of what this check looks for.
- X27 Collection changed while being enumerated — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. This repository's C# was parsed (it ships no .sln or .csproj, so it was read as syntax-only projects), but this check proves its findings from resolved symbols and a reference-free parse resolves none. It abstained rather than report a clean bill it could not earn. That is a gap in this analyzer's reach into build-less repositories — not a finding that the repository is free of what this check looks for.
- X3 Exception handling — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. This check is implemented over the C# syntax tree, and no C# was loaded on this run, so it had nothing to read. That is a gap in this analyzer's language reach — not a finding that the repository is free of what this check looks for.
- X4 Structured logging — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. This check is implemented over the C# syntax tree, and no C# was loaded on this run, so it had nothing to read. That is a gap in this analyzer's language reach — not a finding that the repository is free of what this check looks for.
Limitations & what we did not check
Watchdog assesses the repository exactly as committed, and only the repository. By design it does not reach outside the source tree: the live cloud account, the running CI/CD pipeline, the host's branch-protection and approval rules, the production configuration, or a restore actually exercised against a backup are all out of scope. That boundary is a feature, not a gap — a repo-relative, deterministic scan re-runs identically on any commit and every finding opens at a real file and line, where a live audit can neither be reproduced nor traced. The visible consequence is that controls which leave no in-repo evidence are reported as "not evidenced" and excluded from the score rather than awarded a number a static scan cannot justify.
Per-dimension blind spots
- D1 Cyclomatic Complexity: Cyclomatic complexity counts branches statically — it cannot tell an essential decision tree from accidental tangle, nor see complexity that lives in data or configuration (large switch-case token tables, DSL lexers/parsers, data-as-code rule tables) rather than control flow: a tokenizer's many single-character cases read as high complexity though each branch is trivial.
- D2 Cognitive Complexity: Cognitive-complexity heuristics approximate how hard code is to follow; genuine domain difficulty and well-named intent that eases reading are not captured.
- D3 God Classes: "God class" is sized by members and responsibilities visible in the type — a deliberately broad facade over a coherent subsystem can read the same as an accidental grab-bag. For front-end JS the file-length check is cohesion-aware (a single-responsibility module — one class/IIFE — earns a 3× threshold), but cohesion is approximated from top-level declarations, not true dependency structure.
- D4 Code Duplication: Duplication is token-similarity — an in-process token-stream comparison over sliding windows, with type-aware normalization — so it finds copy-paste, not semantic duplication expressed differently. Committed machine-written code (scaffolded migrations, designer/codegen output, protobuf/OpenAPI stubs, model snapshots) is EXCLUDED — its repetition is the tool's, not the team's — so the score reflects hand-written duplication only.
- D5 Coupling: Coupling is measured between projects/assemblies — runtime coupling through DI, reflection, messaging or shared databases is invisible to a static reference graph.
- D6 Cohesion (LCOM4): LCOM4 cohesion is syntactic — it infers connectivity from which methods touch which fields/methods by name, not from real runtime behaviour or intent.
- D9 Test Distribution: The test-pyramid shape is inferred from project/folder naming and references, with a single test host bucketed per-file by its path tier and content signals — a suite that names tiers unconventionally and gives no per-file signal can still be mis-bucketed.
- D13 Secret Scanning: Secret detection is signature- and entropy-based on the current tree — a secret that does not match a known pattern, or one already rotated, will not be flagged (a clean scan is "nothing matched", not "no secrets exist").
- D14 License Compliance: License compatibility is checked against declared package metadata and a policy — mislabelled or missing license metadata, and obligations that depend on how you distribute, are not resolved here.
- D15 Churn × Complexity Hotspots: Churn hotspots come from git history — a freshly imported or squashed repository has no churn signal, and recent rewrites can mask a historically risky file.
- D16 Bus Factor: Bus-factor is a time-decayed model of commit attribution (who has recently, repeatedly worked a file), not comprehension — pairing, review and reading-without-committing spread knowledge it can't see; bot commits and shared accounts still distort it.
- D20 ADR Quality: ADR quality is an LLM read of the decision records present — it cannot know about decisions made and never recorded, and its verdict is sampled and advisory.
- D26 Project Cohesion: Project focus is sized from members/namespaces per project — a project that is broad by deliberate design reads the same as one that has sprawled.
- D28 Secrets (history): Secret-history scanning sweeps the git log for known patterns — a secret that predates the available history, or never matched a signature, is not found (clean means "nothing matched in the history we can see").
- D29 Static Analysis (SAST): SAST findings are pattern-based (semgrep) — it finds classes of bug it has rules for; logic flaws, auth/authorization gaps and issues needing runtime context are out of reach (and clean means "no rule matched").
- D30 Dependency Vulnerabilities: CVE matching depends on accurate package/version metadata and on the advisory databases — a vulnerability with no published advisory, or in code not declared as a dependency, is not seen. Coverage needs a RESOLVED graph: an unpinned requirements.txt, or a pom without a resolved build, yields partial coverage rather than a clean verdict. An ecosystem the analyzer cannot scan is reported as unmeasured, never as clean.
- D31 IaC & Container Security: IaC scanning checks Dockerfiles/Terraform/Kubernetes against best-practice rules — it cannot see the live cloud account, runtime configuration, or drift between the committed config and what is actually deployed.
- D34 Knowledge Freshness: Freshness is decayed commit RECENCY, not comprehension — code read often but rarely committed reads as orphaned, and stable code that genuinely needs no changes is penalised the same as forgotten code; bot/squash commits distort it like the bus factor.
- D35 Change Coupling: Change coupling is co-change in COMMITS — files split across separate commits, or coupled only through a shared config/build step, read as uncoupled, and a sweeping commit (rename/format) is excluded so it doesn't couple everything. It shows that files change together, not WHY: a high coupling can be a healthy cohesive pair as readily as a hidden leak.
- D43 Malicious Dependencies: Only packages some vulnerability database has already NAMED as malicious are seen — a compromise published in the last hours, or never reported at all, is invisible here, and this dimension reading 10 is not evidence that a dependency is trustworthy. There is no typosquat or dependency-confusion analysis: a package nobody has reported is simply absent from the feeds. Coverage is the dependency scan's: an ecosystem that could not be scanned is disclosed as unmeasured, never as clean.
- AX10 Code composition: Role is inferred from namespace/folder convention, not semantics — a domain concept living in a folder named "Services" reads as application, and the split is lines-of-code, not business value. The business-logic-share score is a SOFT, FLOORED signal: it contributes to the Architecture lens but is floored at the Critical gate, so an infrastructure-heavy design (a gateway, an ETL, a driver) is legitimately low without being nuked to zero.
- AX9 CQS / query purity: Handlers are found by interface/name convention — a query handler using neither is not seen. Mutation is a resolved write/publish invocation (SaveChanges/repository/bus), so a write hidden behind a hand-rolled wrapper, reflection, or a string-keyed service locator resolves to a non-persistence type and isn't flagged; it detects that a query writes state, not whether the write is a legitimate read-side cache update. Clean means "no resolved write/publish in a query body", not a proof of CQS purity.
- DM4 Rich vs anemic domain model: Behaviour is detected as state mutation inside a method body — a method that enforces an invariant by validating-and-throwing without mutating reads as a query, and mutation delegated through an interface the scan can't resolve isn't credited; entities with zero public properties still drop out of the population. It detects that state changes, not whether the rule is correct.
- DM6 Domain ↔ infrastructure boundary: Infrastructure reached through a hand-rolled wrapper, a domain-named facade, reflection, or a string-keyed service locator resolves to a non-infra type and isn't seen; the body scan is symbol resolution over syntax, not full dataflow. A clean result means "no resolved infra reference in a domain body", not a proof of purity.
- M4 Documentation accuracy: Onboarding quality is an LLM read of the docs/setup present — it cannot run the onboarding or measure how long a real new joiner takes; the verdict is sampled and advisory.
- P4 Deployment & Rollback: Approval/branch-protection rules live in repository settings the scan cannot see — only their in-repo evidence (config files, workflows) is checked, so a control enforced purely in the host's settings reads as "not evidenced".
Dimensions
D1 · Cyclomatic Complexity
12 method(s) exceeded the cyclomatic complexity threshold of 15; the worst was ptx::pass::replace_instructions_with_functions::run_instruction at 60. A further 21 method(s) were over the threshold but excluded as flat dispatchers (a long switch/match over independent cases: many branches, almost no nesting), the largest being MethodEmitContext::emit_instruction at 65 — they are counted neither in the figure above nor in this dimension's score. 10 files carry no cyclomatic complexity row at all for this reason — every one of their over-threshold functions was excluded, so the exclusion is disclosed nowhere in the file itself: zluda/src/impl/device.rs (zluda::impl::device::get_attribute at 56), zluda_dnn/src/impl.rs (zluda_dnn::impl::dnn9::get_error_string at 49), zluda64_server/src/main.rs (zluda64_server::main at 44), zluda/src/impl/driver.rs (zluda::impl::driver::get_error_string at 36), format/src/dnn9.rs (format::dnn9::cudnn9_print_elements at 32), and 5 more not listed here. They are named here because the per-file figures other dimensions report are taken BEFORE this exclusion, so such a file can show a high maximum complexity elsewhere in this report and nothing here, with nothing to reconcile the two.
What to do
- Bring the 1 body over 30 down to 30 or less in Cyclomatic Complexity — start with ptx::pass::replace_instructions_with_functions::run_instruction (cyclomatic 60). — Refactoring it lifts Cyclomatic Complexity from 7.6 to about 8.3/10.
- Enforce Cyclomatic Complexity in CI to reach Verified (currently Documented). — Hardens enforcement from Documented toward Prevented — provenance only; does not change the score.
D2 · Cognitive Complexity
28 method(s) exceeded the cognitive complexity threshold of 15; the worst was zluda_bindgen::cuda_derive_display_trait_for_item at 62.
What to do
- Bring the 28 bodies over 15 down to 15 or less in Cognitive Complexity — start with zluda_bindgen::cuda_derive_display_trait_for_item (cognitive 62), OpcodeDefinitions::new (cognitive 58), dark_api::cuda::guid_to_name (cognitive 46). — This score is capped by its worst body, so a finding fixed alone moves it by almost nothing — the next one down takes its place. Refactoring these 28 together lifts Cognitive Complexity from 6.5 to about 10.0/10, projected with the scoring formula itself and assuming each lands exactly at 15.
- Enforce Cognitive Complexity in CI to reach Verified (currently Documented). — Hardens enforcement from Documented toward Prevented — provenance only; does not change the score.
D3 · God Classes
37 god class(es) detected.
What to do
- Resolve the 19 FileTooLong finding(s) in God Classes — start with lib.rs (10), main.rs (2), mod.rs (2). — One of this dimension's main actionable groups (19 warning-level).
- Resolve the 12 FunctionTooLong finding(s) in God Classes — start with replace_instructions_with_functions.rs (2), mod.rs (2), main.rs. — One of this dimension's main actionable groups (12 warning-level).
- Resolve the 2 TooManyFunctions finding(s) in God Classes — start with impl.rs, memory.rs. — One of this dimension's main actionable groups (2 warning-level).
- Enforce God Classes in CI to reach Verified (currently Documented). — Hardens enforcement from Documented toward Prevented — provenance only; does not change the score.
D4 · Code Duplication
67 duplicated block group(s) detected. A further 4 rows report members as variants of one another; they aggregate block groups already counted above and are not themselves counted.
What to do
- Resolve the 6 Duplicated block (8 lines × 2) finding(s) in Code Duplication — start with lib.rs (3), win.rs, driver.rs. — One of this dimension's main actionable groups (6 warning-level).
- Resolve the 6 Duplicated block (6 lines × 2) finding(s) in Code Duplication — start with lib.rs (4), normalize_identifiers.rs, driver.rs. — One of this dimension's main actionable groups (6 warning-level).
- Resolve the 5 Duplicated block (9 lines × 2) finding(s) in Code Duplication — start with deparamize_functions.rs (2), driver.rs, main.rs. — One of this dimension's main actionable groups (5 warning-level).
- Enforce Code Duplication in CI to reach Verified (currently Documented). — Hardens enforcement from Documented toward Verified — provenance only; does not change the score.
D5 · Coupling
46 production modules (Cargo), 0 dependency cycle(s), 0 unstable depended-on module(s). Read from the build's own module declarations; 8 module(s) off the main sequence, with abstractness counted on 34 of the 46 (the rest declare no modelled class or interface, export only macros, declare types in source no model reads, or have no source directory of their own).
What to do
- Resolve the 7 Off the main sequence finding(s) in Coupling. — One of this dimension's main actionable groups (7 warning-level).
- Enforce Coupling in CI to reach Verified (currently Documented). — Hardens enforcement from Documented toward Prevented — provenance only; does not change the score.
D6 · Cohesion (LCOM4)
0 of 175 classes have LCOM4 above 3.
D9 · Test Distribution
112 test methods: 103 unit, 9 integration, 0 BDD, 0 e2e. The Rust suite contributes 112 `#[test]` function(s) across 21 file(s) declaring at least one; its unit/integration split is Cargo's own — 3 of those file(s) are integration-test targets under a crate's tests/ directory, and the rest are #[test] functions compiled into the crate they test.
D13 · Secret Scanning
Secret scan ran and found no leaked secrets.
D14 · License Compliance
0 of 444 shipped crate(s) use a banned license. Licences were resolved from crates.io over the crates a consumer compiles — this repository's REDACTED closed over its manifests' `[dependencies]` and `[build-dependencies]`. Crates it asks for ONLY under `[dev-dependencies]` are excluded: they are not compiled by anything that depends on this repository.
D15 · Churn × Complexity Hotspots
Top hotspots: ptx/src/pass/llvm/emit.rs (6×65=390); ptx/src/pass/replace_instructions_with_functions.rs (4×60=240); ptx/src/pass/mod.rs (2×18=36)
What to do
- Resolve the 3 Hotspot finding(s) in Churn × Complexity Hotspots — start with emit.rs, replace_instructions_with_functions.rs, mod.rs. — One of this dimension's main actionable groups (3 warning-level).
D16 · Bus Factor
11 source file(s) have their living knowledge concentrated in one author (≥90% of recent, decayed contribution). The largest is zluda/src/impl/tex.rs. Counted over 99 of the 136 production source files in this repository: the rest are under the ~2,400-byte size floor this dimension measures over.
What to do
- Resolve the 1 Off-boarding risk finding(s) in Bus Factor. — One of this dimension's main actionable groups (1 recommendation-level).
D20 · ADR Quality
No architecture decision records were found.
What to do
- Resolve the 1 No ADRs found finding(s) in ADR Quality. — One of this dimension's main actionable groups (1 recommendation-level).
D26 · Project Cohesion
1 of 47 build units (Cargo) flagged as possibly oversized/incoherent.
D28 · Secrets (history)
gitleaks scanned the full history AND the current working tree and found no secrets.
D29 · Static Analysis (SAST)
45 finding(s): 0 critical, 45 high, 0 medium, 0 low. 40 unpinned-GitHub-Actions row(s) are reported here but scored by D36 (supply-chain provenance), which measures that posture as `pinned_actions` — one pinning decision is charged once, not once per lens. semgrep hit a parse error in 2 file(s) — `format/src/lib.rs` (line 14, line 23, line 34, …), `zluda_ld/src/lib.rs` (line 7, line 12) — so no absence of findings in the named regions is evidence of anything; rows reported elsewhere in those files are real. Fix the syntax error (or exclude the file deliberately) and re-scan to cover them. Separately, one or more rules could not re-parse an embedded snippet in 3 file(s) (e.g. a workflow `run:` block read as shell). Those files WERE scanned and their other rows are unaffected; only those rules' view of those snippets is missing.
What to do
- Resolve the 1 REDACTED finding(s) in Static Analysis (SAST) — start with REDACTED. — One of this dimension's main actionable groups (1 issue-level).
- Resolve the 1 REDACTED finding(s) in Static Analysis (SAST) — start with REDACTED. — One of this dimension's main actionable groups (1 issue-level).
- No action in Static Analysis (SAST) — all 40 REDACTED finding(s) are reported here at file:line but scored by D36 (supply-chain provenance), so none is charged to this dimension. — One of this dimension's main actionable groups (40 issue-level, 0 of them charged here).
D30 · Dependency Vulnerabilities
26 finding(s): 0 critical, 1 high, 25 medium, 0 low.
What to do
- Resolve the 14 Medium advisory (unsound) finding(s) in Dependency Vulnerabilities — start with REDACTED (14). — One of this dimension's main actionable groups (14 warning-level).
- Resolve the 1 High vulnerability finding(s) in Dependency Vulnerabilities — start with REDACTED. — One of this dimension's main actionable groups (1 issue-level).
- Resolve the 8 Medium vulnerability finding(s) in Dependency Vulnerabilities — start with REDACTED (8). — One of this dimension's main actionable groups (8 warning-level).
D31 · IaC & Container Security
4 finding(s): 0 critical, 0 high, 0 medium, 4 low.
D34 · Knowledge Freshness
5 of 99 significant source file(s) are orphaned — their living knowledge has decayed to nothing, so no one currently understands them. The largest is ptx_parser_macros_impl/src/lib.rs. Counted over 99 of the 136 production source files in this repository: the rest are under the ~2,400-byte size floor this dimension measures over.
What to do
- Resolve the 1 Orphaned knowledge finding(s) in Knowledge Freshness — start with lib.rs. — One of this dimension's main actionable groups (1 issue-level).
- Resolve the 4 Orphaned file finding(s) in Knowledge Freshness — start with utils.rs, tests.rs, ffi.rs. — One of this dimension's main actionable groups (4 recommendation-level).
- Resolve the 1 Further orphaned files (smaller) finding(s) in Knowledge Freshness. — One of this dimension's main actionable groups (1 recommendation-level).
D35 · Change Coupling
Strongest change-coupling: lib.rs↔lib.rs 73%; function.rs↔memory.rs 50%
What to do
- Resolve the 1 Boundary-crossing change coupling finding(s) in Change Coupling — start with lib.rs. — One of this dimension's main actionable groups (1 issue-level).
- Resolve the 1 Change coupling finding(s) in Change Coupling — start with function.rs. — One of this dimension's main actionable groups (1 warning-level).
D36 · Supply-chain Provenance & Signing
0/4 supply-chain integrity signals present (provenance, signing, SBOM, pinned actions).
What to do
- Resolve the 3 REDACTED finding(s) in Supply-chain Provenance & Signing — start with REDACTED, REDACTED, REDACTED. — One of this dimension's main actionable groups (3 warning-level).
- Resolve the 1 REDACTED finding(s) in Supply-chain Provenance & Signing. — One of this dimension's main actionable groups (1 warning-level).
- Resolve the 1 REDACTED finding(s) in Supply-chain Provenance & Signing — start with REDACTED. — One of this dimension's main actionable groups (1 warning-level).
D43 · Malicious Dependencies
No dependency in any ecosystem this repository declares is published as malicious.
Frontend & cross-cutting dimensions
AX10 · Code composition
What to do
- The domain core is a small share of production code, but most of the rest matched no layer vocabulary at all — so this is not yet an anemic-domain finding. The namespace/path convention could not place that code, which makes the composition above a statement about the naming, not about the design. Name the layers (or check that the repository's conventions differ from the ones this check knows) before reading a thin domain into it.
AX3 · Project dependency cycles
AX4 · Dependency direction
AX9 · CQS / query purity
DM4 · Rich vs anemic domain model
DM5 · Encapsulated state
DM6 · Domain ↔ infrastructure boundary
GD1 · Unfinished & placeholder code
M1 · Documentation (README)
- The root README is 42 words, against a bar of 120. Of the three newcomer-critical sections this check looks for by heading, it found no a build/run or getting-started section, no a testing section, no an architecture or project-map section. Sections are matched on HEADING text only, so material written under a heading this check does not recognise — or with no heading at all — is not seen and this row may understate what the document covers.
- 90 code files changed in the last 6 months but the README was not touched — it may no longer reflect the system.
What to do
- Expand the README with getting-started, architecture overview and a project map.
- Add a build/run (quick start) section to the root README — the first thing a newcomer needs.
- Explain in the README how to run the tests. Add a section called Testing to the README in the repository's top folder, saying how to run the test suite.
- Add an 'Architecture' / 'How it works' section to the root README — the high-level shape.
- Add a README to the 52 of 54 project(s) that lack one — worth up to 1.9 pts.
- Review the README against recent changes; refresh the parts that drifted.
M2 · Architecture documentation
- No Architecture Decision Records found — no conventional ADR directory, no numbered `NNNN-title` documents in any markup this check reads, and nothing ADR-shaped by content. Design rationale recorded elsewhere (a design-notes tree, a mailing list, pull-request discussion) is not visible to this check and is not re-findable per decision, so a future maintainer cannot ask why one choice was made and get an answer.
What to do
- Record your significant design decisions. Write one short document per decision, giving its date, the situation, the decision and its consequences, and keep them together with your other design documents. A docs/adr folder with each file named by number and title is the easiest place to find them.
M3 · Folder & project structure
M4 · Documentation accuracy
P1 · CI/CD gates
P10 · Library API & versioning
- 1327/1436 types (92%) declared in the published library are public. For a library, every public type is a stability contract — keep implementation types off the surface and expose only the intended API.
What to do
- Keep implementation types off the public surface (Rust: `pub(crate)`, or keep the module private) so internals can change without breaking consumers.
P2 · Observability
- Only 1/10 runnable modules use logging (modules with no entry point or server are excluded — they are libraries a runnable module hosts). Silent: `.`, `compiler`, `cuda_check`, `ptxas`, `xtask` and 4 more.
What to do
- Extend structured logging to the remaining runnable modules so everything you run is diagnosable in production.
P3 · Security & performance tooling
What to do
- Enable Dependabot/Renovate or a dependency-review gate.
- Add gitleaks/trufflehog in CI to block PRs that introduce committed secrets.
P4 · Deployment & Rollback
- The release is automated and no gate that pauses it for a human is DECLARED IN THIS REPOSITORY'S PIPELINE FILES. What was read: every file under `.github/workflows/`, `.forgejo/workflows/`, `.gitea/workflows/`, `.azuredevops/` and `.azure-pipelines/`, plus `.gitlab-ci*` and `azure-pipelines*` — with comment text stripped, so documenting a gate is not declaring one. What would have counted: GitLab's `when: manual`, CircleCI's `type: approval`, an Azure `ManualValidation@` task or an `approvals:` block, a Jenkins `input` step, a `uses:` step naming an approval action, an `environment:` paired with `reviewers` / `required_reviewers` / `protection` / `wait-timer` / `deployment_branch_policy`, a draft-release step, a `workflow_dispatch` promotion, or a release-event gate. ★ What this cannot see, because none of it is a file: a GitHub environment whose required reviewers are configured in repo SETTINGS, a branch protection rule, or an organisation deployment policy — all of them real, enforced gates that live outside the repository. If yours is one of those, this row is wrong and nothing in the tree could have told us. Otherwise: whatever the release trigger points at is published to users unreviewed, so a mistagged or unverified commit ships and the only remedy is a follow-up release.
What to do
- Nothing pauses a release for a human: publish as a draft release (or gate the release job on a protected tag/manual dispatch) so a bad build can be stopped before users can download it.
P8 · Schema migrations
X12 · Unreachable branch
X13 · Undrained process stream
X14 · Bypassable address classification
X15 · Unvalidated length from an untrusted reader
X16 · Unfloored truncation loop
X17 · Uncapped recursion over a caller-supplied document
X18 · Disposal-pattern correctness
X19 · Unrestored process-global state
X20 · Mistyped argument guard
X21 · Side-effecting pattern guard
X22 · Contradicted release guard
X23 · Unguarded diagnostic materialisation
X24 · Document value interpolated into markup unescaped
X25 · Inert configuration knob
X26 · Unsynchronised callback handoff
X28 · Index access outside its own emptiness guard
X29 · Per-element action decided by a fixed element
X30 · Support guard that admits what it rejects
X32 · Type resolved by simple name across every loaded assembly
X5 · Nullable reference types
- 0/1 NRT-eligible project(s) enable <Nullable>enable</Nullable> (projects targeting a pre-C#-8 framework are excluded — NRTs aren't available there). NRTs catch a whole class of null-deref bugs at compile time.
What to do
- Enable <Nullable>enable</Nullable> across all projects and resolve warnings rather than suppressing with `!`.
Reference — by lens
| Lens | Score | Rating | Impact |
|---|---|---|---|
| Code Health | 46% | Weak — gated by X5 | Capped at Fair by a Critical contributor — resolve it before relying on this lens. |
| Architecture | 98% | Exemplary | Solid. |
| Maturity | 54% | Adequate — gated by M1, M2 | Capped at Fair by a Critical contributor — resolve it before relying on this lens. |
| Readiness | 59% | Adequate | Acceptable, with room to improve. |
| Security | 65% | Adequate — gated by D36 | Capped at Fair by a Critical contributor — resolve it before relying on this lens. |
| Domain Modelling | 100% | Exemplary | Strongest area. |
Unscored — 1 check(s) recorded observations but carry no score
- SC1 Supply-chain hygiene — 1 observation(s) recorded · Advisory — this card reports evidence and never carries a score, so there is nothing missing here.
Not evidenced — 2 control(s) we could not find positive evidence for
- P5 DR & Backup — not evidenced — repo shows no backup/RTO/RPO controls; absence of evidence is not evidence of a working control
- P6 Release Hygiene — not evidenced — no changelog, version stamp or semver release tag in the repo
Not included — 65 check(s) not relevant to this codebase
- AC1 Text alternatives — No web markup found — accessibility is not applicable to this repository.
- AC2 Forms & labels — No web markup found — accessibility is not applicable to this repository.
- AC3 Page structure — No web markup found — accessibility is not applicable to this repository.
- AC4 Keyboard semantics — No web markup found — accessibility is not applicable to this repository.
- AC5 ARIA correctness — No web markup found — accessibility is not applicable to this repository.
- AC6 Visual & motion safety — No web markup found — accessibility is not applicable to this repository.
- AC7 A11y enforcement — No web markup found — accessibility is not applicable to this repository.
- AX1 Captive dependencies — no DI registrations detected
- AX2 Stateful singletons — no singleton implementations detected
- AX5 Architecture & structure — not assessed — architecture style/structure is computed from a project graph (projects, types, module namespaces) that was not loaded for this repository, because the repository is written in a language this check does not yet model, or because its projects failed to load. This is a gap in the analyzer, not a finding about this repository
- AX6 Interface segregation — no public interfaces
- AX7 Slice cohesion — not applicable — not a vertical-slice architecture
- AX8 Test isolation — no test/production split to check
- AXB2 Runtime readiness — Advisory — this card reports evidence and never carries a score, so there is nothing missing here.
- AXR1 Runtime accessibility — docker build failed (exit 1) — -------------------- ERROR: failed to solve: nvidia/cuda:13.0.1-base-ubuntu24.04: failed to resolve source metadata for docker.io/nvidia/cuda:13.0.1-base-ubuntu24.04: failed to do request: Head "https://registry-1.docker.io/v2/nvidia/cuda/manifests/13.0.1-base-ubuntu24.04": Forbidden; runtime evidence skipped This is a statement about this run, not a statement about your application: nothing here says the surface is inaccessible, only that it was never rendered.
- C1 Data Protection — No personal data detected in the analyzed source — no PII-typed entity/column names (Email, FirstName, DateOfBirth, …), no ASP.NET Identity / user-account model, and no stored user credentials. GDPR data-protection controls are therefore N/A here. If this is intentional, record the no-PII posture in an ADR; if the app does process personal data, name those fields conventionally so this dimension activates.
- C2 Access Controls — No access-control surface detected in the analyzed source — no web/app surface to authorize (no HTTP API or web-UI project) and no authorization code at all (no [Authorize]/policies, no imperative guard methods). Access control is therefore N/A here — this is a library/CLI, which is authorized by its CALLER, not by itself. If this codebase grows request handlers, the dimension reactivates and a default-deny posture is expected then.
- C3 Audit Trail — No personal data detected in the analyzed source — no PII-typed entity/column names (Email, FirstName, DateOfBirth, …), no ASP.NET Identity / user-account model, and no stored user credentials. GDPR data-protection controls are therefore N/A here. If this is intentional, record the no-PII posture in an ADR; if the app does process personal data, name those fields conventionally so this dimension activates.
- C4 Data Retention — No personal data detected in the analyzed source — no PII-typed entity/column names (Email, FirstName, DateOfBirth, …), no ASP.NET Identity / user-account model, and no stored user credentials. GDPR data-protection controls are therefore N/A here. If this is intentional, record the no-PII posture in an ADR; if the app does process personal data, name those fields conventionally so this dimension activates.
- C5 Data-Subject Rights — No personal data detected in the analyzed source — no PII-typed entity/column names (Email, FirstName, DateOfBirth, …), no ASP.NET Identity / user-account model, and no stored user credentials. GDPR data-protection controls are therefore N/A here. If this is intentional, record the no-PII posture in an ADR; if the app does process personal data, name those fields conventionally so this dimension activates.
- D10 Test Quality — ~2902 lines of test source are present (.rs) but no test cases reached the test census for this repository — no test root we could resolve declared them, or the files we read declare no cases in a test framework we recognise — so skipped/assertion-free tests couldn't be counted. Not scored — this is a gap in the analyzer, not a finding about this repository.
- D11 Test Reliability — Test reliability not included — the .rs suite was found but not re-run
- D12 Dependency Hygiene — Dependency hygiene not measured
- D17 Explicit Debt — the C# workspace loaded 0 projects, so explicit-debt density could not be measured
- D18 Solution Shape — D18 scores the shape of a C#/VB .NET solution; this repository's .NET projects are all F# (.fsproj), which the C#/VB workspace does not load, so the dimension does not apply.
- D19 Documentation Quality — D19 reads .NET projects as units only; 47 folder(s) declared by Cargo.toml were not judged for an overview, an example or the map.
- D21 Naming Consistency — Naming consistency is measured on C# only; Python, Rust not measured.
- D22 Internal API Consistency — API consistency is measured on C# only; the exposed Rust surface was not measured.
- D23 Boundary Type-Coupling — No bounded-context organisation was detected either — neither a context-shaped layout, nor 2+ sibling source directories each declaring an aggregate root, nor 2+ projects each declaring an entity of their own. Declaring this codebase's bounded contexts (≥2) would let cross-boundary type coupling be assessed. Declare them in `.codehealth/config.yaml` at the repository root (create it if absent), mapping each context name to the module-path or namespace prefixes that belong to it — e.g. `architecture:` → `contexts:` → `Billing: ["src/billing", "Acme.Billing"]`, `Catalog: ["src/catalog", "Acme.Catalog"]`.
- D24 Comment Value — D24 reads C# and Visual Basic comments only — this repository's production source is .py, .rs, which was left unread. Not scored: this is a gap in the analyzer, not a verdict about this repository.
- D25 ADR Conformance — no ADRs to check
- D27 Navigability — symbol resolution incomplete — too few calls resolved to assess navigability
- D32 Data Compliance (PII/GDPR) — No personal data was found crossing a boundary the PII/GDPR ruleset checks — nothing written to a log or console sink, placed in a URL or query string, or persisted to browser storage. That is a clean result for the LEAK surface only: this ruleset detects personal data escaping, it does not inventory the personal data a repository holds, so it is not evidence that this repository has no personal-data surface. The personal-data map (Appendix C) and the C1-C5 compliance cards are what speak to that.
- D37 Vulnerability-disclosure Policy — No vulnerability-disclosure policy file found (SECURITY.md/.markdown/.rst/.txt at root or under .github/.forgejo/.gitea/docs, .well-known/security.txt). A coordinated-disclosure policy may live off-repo, so this is not evidenced rather than failed.
- D39 IL Efficiency — D39 measures the IL emitted by a .NET build; this repository has no .NET solution or project files, so the dimension does not apply.
- D40 Network Egress Confinement — No Kubernetes/orchestration workloads found in the repository manifests; network egress policy is a cluster-native control that may live at the platform/firewall layer, so there is nothing to assess here.
- D41 Kernel & Syscall Confinement — No Kubernetes/orchestration workloads found in the repository manifests; seccomp/AppArmor/SELinux confinement is a workload-level control, so there is nothing to assess here.
- D42 Runtime Threat Enforcement — No Kubernetes/orchestration workloads found in the repository manifests; runtime threat-detection and admission-control policy are cluster-level controls, so there is nothing to assess here.
- D44 Platform End-of-Life — Platform end-of-life not assessed — this repository declares no platform this pass reads
- D7 Architectural Integrity — no checkable ADRs, and no project-reference graph for the cycle pass to read — so this dimension makes no claim about dependency cycles in either direction (where this repository's language has an import-cycle lens, cycles are reported there). Architectural integrity not assessed
- D8 Code Coverage — Coverage not included — suite not readable by the collector
- DM1 Aggregate boundaries — this Rust crate declares no domain aggregate (a struct with private state and its own command methods), so DM1's aggregate-boundary read has no population to be taken over
- DM2 Strongly-typed ids — the crate defines newtype typed-ids but declares no domain aggregate that could carry an id — so DM2's adoption ratio has no population to be taken over
- DM3 Integration-event coupling — not applicable: no DDD domain layer was detected (no aggregate, entity or value-object structure), so there is no domain model for this card to grade
- DM7 Repository granularity — not applicable: no DDD domain layer was detected (no aggregate, entity or value-object structure), so there is no domain model for this card to grade
- ED2 Event/command shape — not scored — deciding whether a command has more than one competing handler requires resolving the call graph, and a call made through an inferred or generic receiver has no resolvable owner in the source. Reported as guidance rather than measured
- ED5 Idempotency — no mutating command handlers or message consumers detected — idempotency check not applicable (this repository commits no MSBuild project, so its C# was read as syntax-only projects — a handler marked ONLY by an interface from a package is not visible without a build)
- ES1 Event Sourcing — not scored — this repository shows none of the 3 signals this lens looks for
- IC1 Incompleteness & stubs — not analysed — these correctness checks read a source model that was not loaded for this repository, because the repository is written in a language this check does not yet model, or because its projects failed to load. This is a gap in the analyzer, not a finding about this repository
- P12 CI test-gate honesty — Reported, not scored — and nothing was matched here. The coverage check applies to any stack, and the automatic-re-run check to any GitHub-Actions workflow, but the checks for excluded tests, skipped tests and sleep-based synchronisation currently recognise only some ecosystems' test-runner idioms, so on a repository built with another stack the zeros below mean 'not checked', not 'clean'.
- P7 Outbound HTTP resilience — not applicable — no HTTP server, API framework or worker entry point was found in the Python, Rust source, so there is no service whose uptime a failing dependency could take down
- P9 Domain vs controller coverage — no coverage report found on disk — produce a coverage report in a standard format (lcov — `cargo install cargo-llvm-cov`, then `cargo llvm-cov --lcov --output-path lcov.info`) and commit it — a hosted scan measures a clone of the repository, so a report that exists only in a working tree, a CI runner's or your own, never reaches it; the artefact is commonly gitignored, so `git add -f` that one file (or un-ignore its path) and commit it alongside the code it measures, or wire coverage collection into CI, to enable this cross-layer check
- PF1 Benchmark discipline — Performance is assessed only for perf-relevant repos — a packaged library, one that ships benchmarks, or one already using allocation-aware APIs. This repo isn't one, so the Performance lens is not applicable and is excluded from the score.
- PF2 Allocation hygiene — Performance is assessed only for perf-relevant repos — a packaged library, one that ships benchmarks, or one already using allocation-aware APIs. This repo isn't one, so the Performance lens is not applicable and is excluded from the score.
- PF3 Async & latency hygiene — Performance is assessed only for perf-relevant repos — a packaged library, one that ships benchmarks, or one already using allocation-aware APIs. This repo isn't one, so the Performance lens is not applicable and is excluded from the score.
- S1 Web-Security Posture — No web surface detected in the analyzed source — no HTTP API or web-UI project (no controllers/minimal-API endpoints, no Razor/Blazor views) and no web middleware (HTTPS redirection, HSTS, security headers, cookies). Transport security, security headers, secure cookies, CSRF/input-validation and middleware-order controls are therefore N/A here — this is a library/CLI/worker, not a web app. Crypto hygiene was still checked and found nothing to flag. If this codebase becomes web-facing, the dimension reactivates automatically.
- X1 Async correctness — not analysed — these correctness checks read a source model that was not loaded for this repository, because the repository is written in a language this check does not yet model, or because its projects failed to load. This is a gap in the analyzer, not a finding about this repository
- X10 Duplicated predicate — Advisory — this card reports evidence and never carries a score, so there is nothing missing here.
- X2 Cancellation propagation — not analysed — these correctness checks read a source model that was not loaded for this repository, because the repository is written in a language this check does not yet model, or because its projects failed to load. This is a gap in the analyzer, not a finding about this repository
- X27 Collection changed while being enumerated — This check needs a compiled C# program and this repository commits no project file to compile, so only its syntax could be read. That is a limit of the analyzer, not a finding about your code.
- X3 Exception handling — not analysed — these correctness checks read a source model that was not loaded for this repository, because the repository is written in a language this check does not yet model, or because its projects failed to load. This is a gap in the analyzer, not a finding about this repository
- X4 Structured logging — not analysed — these correctness checks read a source model that was not loaded for this repository, because the repository is written in a language this check does not yet model, or because its projects failed to load. This is a gap in the analyzer, not a finding about this repository
- X6 Hand-rolled structured-format parsing — Advisory — this card reports evidence and never carries a score, so there is nothing missing here.
- X7 Silent fallback defaults — Advisory — this card reports evidence and never carries a score, so there is nothing missing here.
- X9 Subsumed condition operand — Advisory — this card reports evidence and never carries a score, so there is nothing missing here.
Appendix A — Findings (grouped)
Critical — 48 finding(s)
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- Orphaned knowledge ptx_parser_macros_impl/src/lib.rs:1
- Boundary-crossing change coupling: lib.rs ↔ lib.rs zluda_blas/src/lib.rs:1
Serious — 191 finding(s)
- FileTooLong: llvm/emit.rs ptx/src/pass/llvm/emit.rs
- FileTooLong: src/lib.rs ptx_parser/src/lib.rs
- FileTooLong: src/ast.rs ptx_parser/src/ast.rs
- FileTooLong: src/main.rs zluda_bindgen/src/main.rs
- FileTooLong: instruction_mode_to_global_mode/mod.rs ptx/src/pass/instruction_mode_to_global_mode/mod.rs
- FileTooLong: src/impl.rs zluda_dnn/src/impl.rs
- FileTooLong: src/lib.rs format/src/lib.rs
- FileTooLong: src/lib.rs zluda_trace/src/lib.rs
- FileTooLong: src/lib.rs zluda32/src/lib.rs
- FileTooLong: src/main.rs zluda64_server/src/main.rs
- FileTooLong: pass/mod.rs ptx/src/pass/mod.rs
- FileTooLong: src/lib.rs zluda_common/src/lib.rs
- FileTooLong: src/lib.rs ptx_parser_macros/src/lib.rs
- FileTooLong: src/lib.rs zluda_redirect/src/lib.rs
- FileTooLong: pass/replace_instructions_with_functions.rs ptx/src/pass/replace_instructions_with_functions.rs
- FileTooLong: impl/driver.rs zluda/src/impl/driver.rs
- FileTooLong: src/lib.rs ptx_parser_macros_impl/src/lib.rs
- FileTooLong: src/lib.rs zluda_server_common/src/lib.rs
- FileTooLong: src/lib.rs zluda_windows/src/lib.rs
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- FunctionTooLong: ptx::pass::replace_instructions_with_functions::run_instruction ptx/src/pass/replace_instructions_with_functions.rs:340
- FunctionTooLong: zluda64_server::main zluda64_server/src/main.rs:279
- FunctionTooLong: ptx::pass::insert_post_saturation::run_instruction ptx/src/pass/insert_post_saturation.rs:55
- FunctionTooLong: ptx_parser_macros::emit_definition_parser ptx_parser_macros/src/lib.rs:607
- FunctionTooLong: zluda::impl::device::get_attribute zluda/src/impl/device.rs:69
- FunctionTooLong: ptx::pass::replace_instructions_with_functions::run_statements ptx/src/pass/replace_instructions_with_functions.rs:90
- FunctionTooLong: ptx::pass::instruction_mode_to_global_mode::get_modes ptx/src/pass/instruction_mode_to_global_mode/mod.rs:2127
- FunctionTooLong: format::dnn9::cudnn9_print_elements format/src/dnn9.rs:55
- FunctionTooLong: ptx::pass::replace_instructions_with_functions_fp_required::get_functions ptx/src/pass/replace_instructions_with_functions_fp_required.rs:37
- FunctionTooLong: llvm_zluda::compile::compile llvm_zluda/src/compile.rs:90
- FunctionTooLong: ptx::pass::instruction_mode_to_global_mode::apply_global_mode_controls ptx/src/pass/instruction_mode_to_global_mode/mod.rs:964
- FunctionTooLong: zluda_dnn::impl::run_solution zluda_dnn/src/impl.rs:933
- ptx::pass::replace_instructions_with_functions::run_instruction (cognitive 43) ptx/src/pass/replace_instructions_with_functions.rs:340
- ptx::pass::insert_implicit_conversions::should_convert_relaxed_dst (cognitive 28) ptx/src/pass/insert_implicit_conversions.rs:340
- ptx::pass::normalize_basic_blocks::run (cognitive 26) ptx/src/pass/normalize_basic_blocks.rs:18
- ptx::pass::instruction_mode_to_global_mode::compute_single_mode_insertions (cognitive 22) ptx/src/pass/instruction_mode_to_global_mode/mod.rs:1671
- ptx::pass::remove_unreachable_basic_blocks::run (cognitive 21) ptx/src/pass/remove_unreachable_basic_blocks.rs:9
- ptx::pass::deparamize_functions::run_method (cognitive 20) ptx/src/pass/deparamize_functions.rs:23
- ptx::pass::instruction_mode_to_global_mode::create_control_flow_graph (cognitive 20) ptx/src/pass/instruction_mode_to_global_mode/mod.rs:747
- ptx::pass::insert_implicit_conversions::should_convert_relaxed_src (cognitive 18) ptx/src/pass/insert_implicit_conversions.rs:419
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- Off the main sequence: detours-sys
- Off the main sequence: ptx_parser_macros_impl
- Off the main sequence: zluda_cache
- Off the main sequence: zluda_windows
- Off the main sequence: cuda_types
- Off the main sequence: llvm_zluda
- Off the main sequence: ptx_parser
- Duplicated block (8 lines × 2) cuda_check/src/win.rs:247
- Duplicated block (8 lines × 2) ptx_parser_macros/src/lib.rs:921
- Duplicated block (8 lines × 2) zluda/src/impl/driver.rs:529
- Duplicated block (8 lines × 2) zluda32/src/lib.rs:1209
- Duplicated block (8 lines × 2) zluda_trace/src/dark_api.rs:98
- Duplicated block (8 lines × 2) format/src/lib.rs:1275
- Duplicated block (6 lines × 2) format/src/lib.rs:1352
- Duplicated block (6 lines × 2) format/src/lib.rs:1476
- Duplicated block (6 lines × 2) ptx/src/pass/normalize_identifiers.rs:177
- Duplicated block (6 lines × 2) zluda/src/impl/driver.rs:343
- Duplicated block (6 lines × 2) format/src/lib.rs:902
- Duplicated block (6 lines × 2) format/src/lib.rs:1378
- Duplicated block (9 lines × 2) ptx/src/pass/deparamize_functions.rs:32
- Duplicated block (9 lines × 2) ptx/src/pass/deparamize_functions.rs:42
- Duplicated block (9 lines × 2) zluda/src/impl/driver.rs:265
- Duplicated block (9 lines × 2) zluda_bindgen/src/main.rs:509
- Duplicated block (9 lines × 2) ptx/src/pass/llvm/emit.rs:4460
- zluda_bindgen::cuda_derive_display_trait_for_item (cyclomatic 30) zluda_bindgen/src/main.rs:2194
- zluda_bindgen::generate_process_address_table (cyclomatic 21) zluda_bindgen/src/main.rs:365
- zluda_bindgen::write_cudnn8_types (cyclomatic 18) zluda_bindgen/src/main.rs:818
- Hotspot: ptx/src/pass/llvm/emit.rs ptx/src/pass/llvm/emit.rs:614
- Hotspot: ptx/src/pass/replace_instructions_with_functions.rs ptx/src/pass/replace_instructions_with_functions.rs:340
- Hotspot: ptx/src/pass/mod.rs ptx/src/pass/mod.rs:363
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- Duplicated block (12 lines × 2) ptx/src/pass/insert_implicit_conversions.rs:344
- Duplicated block (12 lines × 2) zluda/src/impl/driver.rs:204
- Duplicated block (12 lines × 2) ptx/src/pass/llvm/emit.rs:4469
- Duplicated block (10 lines × 3) ptx_parser/src/lib.rs:407
- Duplicated block (10 lines × 3) zluda/src/impl/driver.rs:172
- Duplicated block (10 lines × 3) format/src/dnn9.rs:29
- Duplicated block (7 lines × 2) zluda32/src/ipc.rs:77
- Duplicated block (7 lines × 2) zluda_bindgen/src/main.rs:305
- Duplicated block (7 lines × 2) zluda_dnn/src/impl.rs:1698
- Duplicated block (6 lines × 3) zluda32/src/lib.rs:1212
- Duplicated block (6 lines × 3) ptx/src/pass/deparamize_functions.rs:16
- Duplicated block (6 lines × 3) ptx/src/pass/expand_operands.rs:19
- Duplicated block (5 lines × 2) ptx/src/pass/deparamize_functions.rs:122
- Duplicated block (5 lines × 2) format/src/dnn9.rs:38
- Duplicated block (5 lines × 2) ptx/src/pass/insert_explicit_load_store.rs:104
- ptx::pass::replace_instructions_with_functions::run_instruction (cyclomatic 60) ptx/src/pass/replace_instructions_with_functions.rs:340
- ptx::pass::insert_implicit_conversions::should_convert_relaxed_dst (cyclomatic 17) ptx/src/pass/insert_implicit_conversions.rs:340
- zluda_bindgen::cuda_derive_display_trait_for_item (cognitive 62) zluda_bindgen/src/main.rs:2194
- zluda_bindgen::generate_process_address_table (cognitive 46) zluda_bindgen/src/main.rs:365
- zluda::impl::module::get_best_ptx_and_compile (cognitive 26) zluda/src/impl/module.rs:167
- zluda::impl::surf::convert_resource_desc (cognitive 19) zluda/src/impl/surf.rs:36
- TooManyFunctions: zluda_fft::impl zluda_fft/src/impl.rs:11
- TooManyFunctions: zluda::impl::memory zluda/src/impl/memory.rs:18
- Members sharing a duplicated core (5 members, 50+ identical tokens) nvapi/src/nvapi.rs:14170
- Members sharing a duplicated core (5 members, 50+ identical tokens) zluda_bindgen/src/main.rs:178
- Duplicated block (12–13 lines × 2) zluda/src/impl/graphics.rs:77
- Duplicated block (12–13 lines × 2) zluda_trace/src/lib.rs:1112
- Duplicated block (11 lines × 2) ptx/src/pass/insert_implicit_conversions.rs:36
- Duplicated block (11 lines × 2) zluda_bindgen/src/main.rs:788
- Duplicated block (10 lines × 2) ptx/src/pass/instruction_mode_to_global_mode/mod.rs:1174
- Duplicated block (10 lines × 2) ptx/src/pass/llvm/emit.rs:2527
- Duplicated block (7 lines × 3) zluda_bindgen/src/main.rs:187
- Duplicated block (7 lines × 3) zluda_bindgen/src/main.rs:217
- Duplicated block (5 lines × 3) format/src/lib.rs:1437
- Duplicated block (5 lines × 3) ptx/src/pass/insert_explicit_load_store.rs:93
- dark_api::cuda::guid_to_name (cyclomatic 29) dark_api/src/lib.rs:154
- CvtDetails::new (cyclomatic 23) ptx_parser/src/ast.rs:2507
- ptx_parser_macros::emit_definition_parser (cyclomatic 23) ptx_parser_macros/src/lib.rs:607
- OpcodeDefinitions::new (cyclomatic 22) ptx_parser_macros/src/lib.rs:36
- ModuleEmitContext::emit_method (cyclomatic 18) ptx/src/pass/llvm/emit.rs:121
- Statement::visit_map (cyclomatic 18) ptx/src/pass/mod.rs:363
- MethodEmitContext::emit_conversion_default (cyclomatic 17) ptx/src/pass/llvm/emit.rs:838
- OpcodeDefinitions::new (cognitive 58) ptx_parser_macros/src/lib.rs:36
- dark_api::cuda::guid_to_name (cognitive 46) dark_api/src/lib.rs:154
- MethodEmitContext::emit_conversion_default (cognitive 40) ptx/src/pass/llvm/emit.rs:838
- ModuleEmitContext::emit_method (cognitive 32) ptx/src/pass/llvm/emit.rs:121
- Arguments::parse (cognitive 32) ptx_parser_macros_impl/src/parser.rs:525
- ptx_parser_macros::emit_definition_parser (cognitive 29) ptx_parser_macros/src/lib.rs:607
- PartialModeInsertion::new (cognitive 28) ptx/src/pass/instruction_mode_to_global_mode/mod.rs:1888
- CvtDetails::new (cognitive 23) ptx_parser/src/ast.rs:2507
- Zluda32DarkApi::get_module_globals (cognitive 19) zluda/src/impl/driver.rs:516
- CodeLibraryRef::iterate_modules (cognitive 19) zluda_common/src/lib.rs:1217
- DarkApiState2::override_export_table (cognitive 18) zluda_trace/src/dark_api.rs:25
- ptx_parser::multi_variable (cognitive 17) ptx_parser/src/lib.rs:1099
- PropagationState::merge_with_self (cognitive 16) ptx/src/pass/instruction_mode_to_global_mode/mod.rs:1845
- SingleOpcodeDefinition::extract_and_insert (cognitive 16) ptx_parser_macros/src/lib.rs:233
- VariantProperty::parse (cognitive 16) ptx_parser_macros_impl/src/lib.rs:421
- zluda_redirect::is_nvrtc (cognitive 16) zluda_redirect/src/lib.rs:449
- ClassTooLong: MethodEmitContext ptx/src/pass/llvm/emit.rs:476
- TooManyMethods: MethodEmitContext ptx/src/pass/llvm/emit.rs:476
- TooManyFields: hipDeviceProp_tR0600 compiler/src/main.rs:223
- MethodTooLong: ResolvedControlFlowGraph.new ptx/src/pass/instruction_mode_to_global_mode/mod.rs:392
- Change coupling: function.rs ↔ memory.rs zluda/src/impl/function.rs:1
- REDACTED
- REDACTED
- Edited copy of a member (16 corresponding lines) format/src/lib.rs:1436
- Members sharing a duplicated core (59 members, 50+ identical tokens) nvapi/src/nvapi.rs:1356
- Duplicated block (25 lines × 2) zluda/src/impl/driver.rs:640
- Duplicated block (24 lines × 2) ptx_parser_macros_impl/src/lib.rs:657
- Duplicated block (23 lines × 2) zluda_trace/src/lib.rs:327
- Duplicated block (22 lines × 2) zluda/src/impl/surf.rs:56
- Duplicated block (18–21 lines × 2) format/src/dnn9.rs:383
- Duplicated block (20–21 lines × 2) ptx/src/pass/insert_implicit_conversions.rs:369
- Duplicated block (17 lines × 2) ptx_parser/src/lib.rs:444
- Duplicated block (16 lines × 2) ptx/src/pass/llvm/emit.rs:2898
- Duplicated block (15 lines × 3) ptx/src/pass/expand_operands.rs:40
- Duplicated block (14 lines × 5) zluda_bindgen/src/main.rs:202
- Duplicated block (14 lines × 2) zluda_blas/src/impl.rs:95
- Duplicated block (10–13 lines × 5) nvapi/src/nvapi.rs:14171
- Duplicated block (13 lines × 2) ptx/src/pass/instruction_mode_to_global_mode/mod.rs:1184
- Duplicated block (10–12 lines × 3) ptx/src/pass/expand_operands.rs:33
- Duplicated block (11–12 lines × 2) zluda_trace/src/lib.rs:692
- Duplicated block (9–11 lines × 59) nvapi/src/nvapi.rs:1357
- Duplicated block (9–10 lines × 2) format/src/lib.rs:1195
- Duplicated block (9 lines × 59) nvapi/src/nvapi.rs:1367
- Duplicated block (9 lines × 5) nvapi/src/nvapi.rs:14186
- Duplicated block (9 lines × 3) zluda_bindgen/src/main.rs:188
- Duplicated block (8 lines × 3) format/src/dnn9.rs:395
- Duplicated block (6 lines × 6) ptx/src/pass/deparamize_functions.rs:6
- Duplicated block (10 lines × 4) format/src/dnn9.rs:300
- Duplicated block (7 lines × 4) format/src/dnn9.rs:359
- Duplicated block (5 lines × 6) format/src/dnn9.rs:21
- README may be stale
- NuGet dependencies are not locked
Minor — 23 finding(s)
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- Orphaned file llvm_zluda/src/utils.rs:1
- Orphaned file zluda_blas/src/tests.rs:1
- Orphaned file llvm_zluda/src/ffi.rs:1
- Orphaned file ptx/src/pass/llvm/mod.rs:1
- Off-boarding risk: anonymized user #1
- No ADRs found
- Projects may be oversized for their cohesion
- Further orphaned files (smaller)
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- REDACTED
- Thin README
- No ADRs
- Large public API surface
- Logging is not universal
- No release approval gate
- Nullable reference types not enabled everywhere
Minor — 1 finding(s)
- Off the main sequence: kernel_metadata
Appendix B — Reproduction & audit trail
| Dimension | Tool | Version | Command | Findings | Raw output |
|---|---|---|---|---|---|
| D28 · Secrets (history) | gitleaks | — | gitleaks detect --no-banner --config /opt/gitleaks-rules/watchdog-gitleaks.toml --report-format json --report-path /tmp/watchdog-gitleaks-f7efdc2f08f540b483599a2a928db885/history.json --exit-code 0 --source . | 0 | artifacts/raw/gitleaks-history.json |
| D28 · Secrets (history) | gitleaks | — | gitleaks detect --no-git --no-banner --config /opt/gitleaks-rules/watchdog-gitleaks.toml --report-format json --report-path /tmp/watchdog-gitleaks-f7efdc2f08f540b483599a2a928db885/tree.json --exit-code 0 --source . | 0 | artifacts/raw/gitleaks-tree.json |
| D29 · Static Analysis (SAST) | semgrep | — | semgrep --config /opt/semgrep-rules/security-audit.yml --config /opt/semgrep-rules/owasp-top-ten.yml --config /opt/semgrep-rules/watchdog-sast.yml --json --quiet --timeout 10 --timeout-threshold 3 --metrics off --exclude *.arrow --exclude *.avro --exclude *.csv --exclude *.dat --exclude *.db --exclude *.feather --exclude *.geojson --exclude *.h5 --exclude *.hdf5 --exclude *.jsonl --exclude *.log --exclude *.ndjson --exclude *.npy --exclude *.npz --exclude *.orc --exclude *.parquet --exclude *.pickle --exclude *.pkl --exclude *.psv --exclude *.sqlite --exclude *.sqlite3 --exclude *.tfrecord --exclude *.tsv --exclude *.txt --exclude cuda_macros/src --exclude ext/detours/samples --exclude ext/detours/src --exclude ext/detours/tests --exclude ext/detours/vc --exclude ext/highs-sys --exclude ext/hip_runtime-sys --exclude ext/hipblaslt-sys --exclude ext/hipfft-sys --exclude ext/miopen-sys --exclude ext/rocblas-sys --exclude ext/rocm_smi-sys --exclude ext/rocsparse-sys . | 45 | artifacts/raw/semgrep.json |
| D30 · Dependency Vulnerabilities | osv-scanner | — | osv-scanner --format json --recursive --experimental-disable-plugins transitivedependency/requirements --experimental-disable-plugins dotnet/csproj --experimental-disable-plugins dotnet/nugetcpm . /tmp/codehealth-osv-sbom-f7d7587bfeae414ea0acc67378ea42c9 | 26 | artifacts/raw/osv-scanner.json |
| D31 · IaC & Container Security | trivy | — | trivy config --format json --quiet . | 4 | artifacts/raw/trivy-config.json |
| D32 · Data Compliance (PII/GDPR) | semgrep | — | semgrep: not applicable — No personal data was found crossing a boundary the PII/GDPR ruleset checks — nothing written to a log or console sink, placed in a URL or query string, or persisted to browser storage. That is a clean result for the LEAK surface only: this ruleset detects personal data escaping, it does not inventory the personal data a repository holds, so it is not evidence that this repository has no personal-data surface. The personal-data map (Appendix C) and the C1-C5 compliance cards are what speak to that. | 0 | — |
| D37 · Vulnerability-disclosure Policy | disclosure | — | disclosure: not applicable — No vulnerability-disclosure policy file found (SECURITY.md/.markdown/.rst/.txt at root or under .github/.forgejo/.gitea/docs, .well-known/security.txt). A coordinated-disclosure policy may live off-repo, so this is not evidenced rather than failed. | 0 | — |
| D40 · Network Egress Confinement | runtime-hardening | — | runtime-hardening: not applicable — No Kubernetes/orchestration workloads found in the repository manifests; network egress policy is a cluster-native control that may live at the platform/firewall layer, so there is nothing to assess here. | 0 | — |
| D41 · Kernel & Syscall Confinement | runtime-hardening | — | runtime-hardening: not applicable — No Kubernetes/orchestration workloads found in the repository manifests; seccomp/AppArmor/SELinux confinement is a workload-level control, so there is nothing to assess here. | 0 | — |
| D42 · Runtime Threat Enforcement | runtime-hardening | — | runtime-hardening: not applicable — No Kubernetes/orchestration workloads found in the repository manifests; runtime threat-detection and admission-control policy are cluster-level controls, so there is nothing to assess here. | 0 | — |
| D43 · Malicious Dependencies | osv-scanner | — | osv-scanner --format json --recursive --experimental-disable-plugins transitivedependency/requirements --experimental-disable-plugins dotnet/csproj --experimental-disable-plugins dotnet/nugetcpm . /tmp/codehealth-osv-sbom-f7d7587bfeae414ea0acc67378ea42c9 | 0 | artifacts/raw/osv-scanner.json |