PyTorch GitHub commits

Author information

Name: Jagadish Krishnamoorthy
Emails:
 - jagadish.krishnamoorthy@amd.com (recent)
 - jagdish.krishna@gmail.com (older)
Commits till: 05/18/2026

Commit history

Total: 73 commits (70 as primary author, 3 co-authored)

# PR Title Role
1#183633[ROCm] Remove redundant ROCm scale-mode branch in CUDABlasAuthor
2#182195[ROCm] Add CK BLAS backend torch.mm testAuthor
3#180731[ROCm] Run test_scaled_mm_deepseek_error_messages on mi350 archAuthor
4#180518[ROCm] Fix evaluate_platform_supports_fp8 false-positiveAuthor
5#180384[ROCm] Update scaled_mm DeepSeek error messageAuthor
6#179782[ROCm] Enable ROCm swizzle check and update scaled_mm swizzle testsAuthor
7#178570test_scaled_matmul_cuda: Parameterize float8 basics casesAuthor
8#176924Skip 4-GPU distributed tests on 2-GPU runnersAuthor
9#173737[ROCm] Enable scaled group mm on gfx950Author
10#173502[ROCm] group_gemm perf: Add new kernel for all K equal casesAuthor
11#172988CUDAScaledBlas - replace FBGEMM_GENAI with MSLKAuthor
12#172791[ROCm] Unifying hipBLASLt architecture lists into common hook methodsAuthor
13#172311[ROCm] Use HIPCachingAllocator for CK argument and workspace buffersAuthor
14#171901[ROCm] Add unit test to verify grouped GEMM CK opt-in flagAuthor
15#171140[ROCm] Make grouped GEMM CK opt-in via env and default to fallback pathAuthor
16#171121[ROCm] Refactor ROCm CK config generation into shared helperAuthor
17#170254[ROCm] inductor/fp8 test: Check for "cuda" in device typeAuthor
18#169356[ROCm] Enable group gemm on gfx90aAuthor
19#168220[ROCm] Fix HIP document urlAuthor
20#167403[ROCm] Disable group gemm CK path when CK is not enabledAuthor
21#167390[ROCm][CI] trigger magma build with gfx950 for ROCm7.1Author
22#167299[ROCm][CI] unconditionally add gfx950, gfx115x to PYTORCH_ROCM_ARCHAuthor
23#167066[ROCm] Skip nvfp4 tests on ROCmAuthor
24#166478[ROCm][CI] remove relaxed tolerance for tf32 testsCo-author
25#166334[ROCm] Enable group gemm through CKAuthor
26#165968[ROCm] Skip test_blockwise_nvfp4_with_global_scaleAuthor
27#165788test_scaled_matmul_cuda: fix infer_scale_swizzleAuthor
28#165528[ROCm] Add scaled_mm v2 supportAuthor
29#164320MX: Remove redundant PLATFORM_SUPPORTS_MX_GEMM constantAuthor
30#163127[ROCm] Fix mx fp8 and fp4 code after scaling refactor changesAuthor
31#163050[ROCm] test_aot_inductor: Enable fp8 testsAuthor
32#162787[ROCm] Enable test_fixed_stridingAuthor
33#162600[ROCm] rocblas Aten GEMM overload for FP32 output from FP16/BF16 inputsAuthor
34#161540ROCm: Enable overload tests from test_matmul_cudaAuthor
35#161166forward fix of #152198Author
36#161009test_matmul_cuda: Refine MX test skippingAuthor
37#151360AMD/ROCm OCP Micro-scaling Format (mx-fp8/mx-fp4) SupportCo-author
38#150250ROCm: Add trailing comma for consistency in gfx architecture listAuthor
39#148945ROCm: Enable tf32 testing on test_nnAuthor
40#148228ROCm: Disable torch check for Multiplication of two Float8_e5m2 matricesAuthor
41#147186ROCm: Remove static specifier for allow_tf32 variableAuthor
42#144692ROCm: Skip tests in elastic/utils/distributed_testAuthor
43#140319ROCm: Enable 4 gpu tests for distributed configAuthor
44#137724[ROCm] Enable dist sharded_tensor test suitesAuthor
45#92932[ROCm] Fix distributed tests failure and enable ROCm distributed CIAuthor
46#85760[ROCm] [FakeTensorTest] Enable test_fallback_memory_propAuthor
47#84100[ROCm] Enable test_filtering_env_varAuthor
48#84077[ROCm] add case for FP32MatMulPattern skip propertyAuthor
49#82749[ROCm] Enable test_ddp_profiling_torch_profilerAuthor
50#82356[ROCm] Enable test_multiprocessing testsAuthor
51#82005[ROCm] Enable test_grad_layout_1devicemodule_1replicaperprocessAuthor
52#80369[test_shape_ops] Increase system memory requirementAuthor
53#78633[ROCm] Enable test_instantiator, test_type_hintsAuthor
54#78401[ROCm] TestGradients: Enable grad and gradgradAuthor
55#78385[ROCm] enable test_lobpcg_ortho_cuda_float64Author
56#72923[LTS] CherryPick: Add multi gpu checker for TestZeroRedundancyOptimizerCo-author
57#68487[ROCm] Use hipCUB/rocPRIM scan algorithms for large index supportAuthor
58#65203[ROCm] update Magma for 4.3 releaseAuthor
59#62502[ROCm] update magmaAuthor
60#61313[ROCm] Skip test_masked_scatter_large_tensor_cudaAuthor
61#59339[Release-1.9] Disable failing ROCM-4.2 testsCo-author
62#59158[ROCm] disable test_Conv2d_groups_nobias for ROCmAuthor
63#57159[Distributed test] Enable ddp_control_flow tests for ROCmAuthor
64#54714[test_c10d] Add wait in nccl high priority stream testAuthor
65#53850[ROCm] Fix the flaky test test_stream_event_nogilAuthor
66#53564[TestZeroRedundancyOptimizer] Add multi gpu checkerAuthor
67#53192[ROCm] Disable flaky tests on ROCmAuthor
68#51510[ROCm] disable tests for ROCm 4.0.1Author
69#50629[distributed_test_c10d] Enable disabled ROCm testsAuthor
70#50421[distributed_test] Enable disabled ROCm testsAuthor
71#49069[distributed] Provide parameter to pass GPU ID in barrier functionAuthor
72#48023[ROCm] Enable skipped distributed global testsAuthor
73#47898distributed_test: Map rank to GPU accordinglyAuthor

Back to home