[ROCm] Fix Multi-Kernel Support for Triton Templates on ROCm#2817

Closed

chinmaydk99 wants to merge 47 commits intoROCm:developfrom

chinmaydk99:ck-autotune-mismatch

Commits on Nov 18, 2025

Revert "[ROCm] enable fastSpecializedAtomicAdd for gfx950 (#167661)"
pytorchmergebot
committed
[Profiler] Deprecate export_memory_timeline method (#168036)

sraikund16
authored and
pytorchmergebot
committed
`ComplexTensor` subclass (#167621)

hameerabbasi
authored and
pytorchmergebot
committed
[CI] Disable ET tests (again) (#168090)

malfet
authored and
pytorchmergebot
committed
[CUDA][Thor] Enable CUTLASS matmuls on Thor (#164836)

Aidyn-A
authored and
pytorchmergebot
committed
distributed/debug: add an HTTP server for debugging running jobs (#167395)

d4l3k
authored and
pytorchmergebot
committed
[refcycle-logger] Output tensor size in the refcycle visualization (#167079)

czardoz
authored and
pytorchmergebot
committed
[xpu][test] port some distributed tensor test files for Intel GPU (#161703)

wincent8
authored and
pytorchmergebot
committed
small changes (#167852)

eellison
authored and
pytorchmergebot
committed
Fix all gather bucketing fusion in of dtype casts (#167853)

eellison
authored and
pytorchmergebot
committed
Handled erased hiding nodes from dtype bucketing (#167863)

eellison
authored and
pytorchmergebot
committed
[ROCm][CI] Limit caching to ROCm jammy docker images (#168088)

jithunnair-amd
authored and
pytorchmergebot
committed
huber_loss numerical issue (#166952)

jturney
authored and
pytorchmergebot
committed
[codemod][lowrisk] Remove unused exception parameter from caffe2/torch/csrc/Exceptions.h (#168056)

r-barnes
authored and
pytorchmergebot
committed
Fix Tensor use_count check in VariableType.cpp (#168060)

colesbury
authored and
pytorchmergebot
committed
[CI] Migrate all gcc9 jobs to gcc11 (#167933)

malfet
authored and
pytorchmergebot
committed
[MPS] Move `elu` impl to Metal (#166903)

kurtamohler
authored and
pytorchmergebot
committed

Commits on Nov 19, 2025

[ROCm][CI] Update concurrency setting for docker-cache-rocm.yml (#168104)

jithunnair-amd
authored and
pytorchmergebot
committed
typo corrected in type.cpp (#167907)

RajeshvShiyal
authored and
pytorchmergebot
committed
[ROCm][CI] forward fix libtorch agnostic tests (#168087)

authored and
pytorchmergebot
committed
[CI][CUDA] Unskip nvshmem triton tests (#167760)

nWEIdia
authored and
pytorchmergebot
committed
[xpu][fix] Fix empty cache on mempool (#168074)

guangyey
authored and
pytorchmergebot
committed
Fix stable ABI to/from deprecation warnings. Add my_shape test. (#167923)

pearu
authored and
pytorchmergebot
committed
[2/3][XPU][feature] The implementation of MemPool for XPU (#166833)

majing921201
authored and
pytorchmergebot
committed
[AOTI] Fix a GPU memory leak caused by reference circle (#168063)

desertfire
authored and
pytorchmergebot
committed
[CD] Add `cuda-bindings` dependency to CUDA wheels (#167769)

malfet
authored and
pytorchmergebot
committed
Update AGENTS.md (#168111)

oulgen
authored and
pytorchmergebot
committed
[MPS][1/N] Fix unsupported dtypes error checking for some MPS ops (#166273)

malfet
authored and
pytorchmergebot
committed
Shrink binary size (#168080)

colesbury
authored and
pytorchmergebot
committed
Improve build logic in activities for kineto (#167204)

guangyey
authored and
pytorchmergebot
committed
[vision hash update] update the pinned vision hash (#168130)

pytorchupdatebot
authored and
pytorchmergebot
committed
[hoo] Invoke subgraph + effect (#167231)

angelayi
authored and
pytorchmergebot
committed
[invoke_subgraph] Don't run the graph twice when autograd enabled (#167245)

angelayi
authored and
pytorchmergebot
committed
fix: use grad div factor when fsdp_degree=1 (#167178)

garrett361
authored and
pytorchmergebot
committed
[CUDA][Complex] Bump tolerances for `TestFFTCUDA.test_reference_nd__refs_fft_irfftn_cuda_complex64` (#168016)

eqy
authored and
pytorchmergebot
committed
[DTensor] Fix deadlock after fast cache clear (#168069)

zpcore
authored and
pytorchmergebot
committed
[user-streams] Insert backward syncs (#167747)

mlazos
authored and
pytorchmergebot
committed
[fix] Assign CUDAEvent external member properly (#167711)

guangyey
authored and
pytorchmergebot
committed
LocalTensor for random_ops tests (#166540)

dolpm
authored and
pytorchmergebot
committed
[DeviceMemory] Add Basic Statistics to Device Memory in OpenReg (#166395)

licy666
authored and
pytorchmergebot
committed
Hide all symbols (except stable/headeronly/shim) if TORCH_STABLE_ONLY is defined (#167496)

mikaylagawarecki
authored and
pytorchmergebot
committed
Error when non stable/headeronly/shim headers are included by stable extension (#167855)

mikaylagawarecki
authored and
pytorchmergebot
committed
Revert "[SymmMem] Skip multicast init if any CUDA call fails (#168049)"
pytorchmergebot
committed
Revert "[pytree][compile] Slightly faster TreeSpec init (#168024)"
pytorchmergebot
committed
Fix link for core maintainers request form (#168089)

albanD
authored and
pytorchmergebot
committed
[compile][to_local] Support Sequence-like placement user defined objects (#168149)

anijain2305
authored and
pytorchmergebot
committed
Fixing multi-kernel autotune for different size hints on ROCm

chinmaydk99
authored and
AMD
committed