Commits · 04ef76dd6c36be74bf4d803b6e8dd488bce591e8 · OpenDAS / TransformerEngine

13 Oct, 2025 1 commit
- [DCU] remove channelwise int8 group gemm · 04ef76dd
  yuguo authored Oct 13, 2025
  
  04ef76dd
28 Sep, 2025 1 commit
- bug fix · 650cb815
  dongchl authored Sep 28, 2025
  
  650cb815
24 Sep, 2025 1 commit
- support activation offloading · 162e32d4
  dongcl authored Sep 24, 2025
  
  162e32d4
19 Sep, 2025 2 commits
- support activation offloading · cd4cdf80
  dongcl authored Sep 19, 2025
  
  cd4cdf80
- [DCU] fix · b15412aa
  yuguo authored Sep 19, 2025
  
  b15412aa
18 Sep, 2025 5 commits
- Fix w8a8 lightop restriction · 803be71d
  wenjh authored Sep 18, 2025
```
Signed-off-by: wenjh <wenjh@sugon.com>
```
  803be71d
- Adapt to changes of hipblaslt · d81f8119
  wenjh authored Sep 18, 2025
```
Signed-off-by: wenjh <wenjh@sugon.com>
```
  d81f8119
- Enable lightop w8a8 · 3f800f01
  wenjh authored Sep 18, 2025
```
Signed-off-by: wenjh <wenjh@sugon.com>
```
  3f800f01
- [DCU] fix compile · d55fdf7b
  yuguo authored Sep 18, 2025
  
  d55fdf7b
- [DCU] fix · 00fcd784
  yuguo authored Sep 18, 2025
  
  00fcd784
12 Sep, 2025 1 commit
- Fix std::bad_alloc on test_float8blockwisetensor.py · 4922108e
  wenjh authored Sep 12, 2025
```
Signed-off-by: wenjh <wenjh@sugon.com>
```
  4922108e
11 Sep, 2025 1 commit
- [Perf] blockwise 1d better perf · fca88163
  wenjh authored Sep 11, 2025
```
Signed-off-by: wenjh <wenjh@sugon.com>
```
  fca88163
02 Sep, 2025 1 commit
- Fix build problems while not support fp4 · 8e0fd518
  wenjh authored Sep 02, 2025
```
Signed-off-by: wenjh <wenjh@sugon.com>
```
  8e0fd518
28 Aug, 2025 3 commits

[PyTorch] Disable determinism for sm100 (#2130) · fedd9ddc

Charlene Yang authored Aug 28, 2025



* disable determinism for sm100+ and cudnn<9.14
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* fix remaining CI failures
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci



* revert some changes
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* revert more changes
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* remove sm100 from determinism table
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

---------
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

fedd9ddc

[DCU] fix quantize bug · d86ee4c8
yuguo authored Aug 28, 2025

d86ee4c8
[DCU] fix bugs · 546bb548
yuguo authored Aug 28, 2025

546bb548

27 Aug, 2025 4 commits

Further relax constraints to cuDNN 9.13 for disabling fused attn for kv caching (#2121) · 1d1e8efe
Kshitij Lakhani authored Aug 27, 2025
```
Signed-off-by: Kshitij Lakhani <klakhani@nvidia.com>
```
1d1e8efe

Add cuBLASMp-backed GEMM-like API to TE common (#1824) · d7874aad

Vladimir Cherepanov authored Aug 26, 2025



* Pick up cuBLASMp during build
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Saving...
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Change lib order to fix link error
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Saving...
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Context creation, incomplete...
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Test fixure
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Saving...
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* A sanity AgGemm test, failing...
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Saving...
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Fix axes
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Take care of uneven distribution
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Use MPI to get position of local matrices
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Refactor
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Refactor & fixes
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Saving...
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Gemm-RS
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Gemm-AR, not working...
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Fixes
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Setting all-reduce epilogue for gemm-ar
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Use supported shapes for GEMM-AR
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Tweak tolerance
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* First shot at fp8
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Use TensorHolder in tests
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* More test configs
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Support comm_sm_count
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Parametrize dtypes for A, B and D separately
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Tweak scaling
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Amax ptr
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Flags parity with cublas_gemm, saving...
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Cleanup
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Bias tests
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Fix bias test
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Aux, saving...
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* aux_ld
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* A fix
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Use test::Tensor
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Set scale inv
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Remove unsupported test configs
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Tweak tests
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Replace libcal with NCCL
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Add NVTX markers to API functions
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Tweak GemmAr tests
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* More test config
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Fix merge fallout
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Remove MPI dependency, comment API, add algo parameter
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Fix nvshmem dependency
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Fix nvshmem build
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Excluse CommGemm tests from L0_cppunittest
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Add cpp_distributed sh file for CI
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Adapt tp TensorAllocator
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci



* Skip GemmAr test on unsupported HW
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Oversibscribe is needed on some clusters
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Fix incomplete libcal removal
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Move CI tests to L1
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Rename context to include NVTE prefix
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Remove leftover code
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* NVTE_WITH_CUBLASMP off by default
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* More detailed NVTE_CHECK diag
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci



* Comment API
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Include stdbool header for legacy C compilers
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Remove now unused argument
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Abstract away cuBLASMp algo behind our own enum
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* More detailed shape diag messages
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci



* Update transformer_engine/common/include/transformer_engine/comm_gemm.h
Co-authored-by: Przemyslaw Tredak <ptrendx@gmail.com>
Signed-off-by: Vladimir Cherepanov <56651474+mk-61@users.noreply.github.com>

* Add license
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

---------
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>
Signed-off-by: Vladimir Cherepanov <56651474+mk-61@users.noreply.github.com>
Co-authored-by: Vladimir Cherepanov <vcherepanov@nvidia.com>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Przemyslaw Tredak <ptrendx@gmail.com>

d7874aad

[PyTorch] Add test for TRT integration + fix for mxfp8 export (#2083) · f8d2c50e

Paweł Gadziński authored Aug 20, 2025



* code drop
Signed-off-by: Pawel Gadzinski <pgadzinski@nvidia.com>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci



* fix
Signed-off-by: Pawel Gadzinski <pgadzinski@nvidia.com>

* fix
Signed-off-by: Pawel Gadzinski <pgadzinski@nvidia.com>

---------
Signed-off-by: Pawel Gadzinski <pgadzinski@nvidia.com>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

f8d2c50e

[DCU] fix compile · 5b6190b2
yuguo authored Aug 27, 2025

5b6190b2

26 Aug, 2025 11 commits

Revert "[Common] PDL for Blockwise Quantization" (#2115) · 58c3ac80

jberchtold-nvidia authored Aug 26, 2025

Revert "[Common] PDL for Blockwise Quantization (#2066)"

This reverts commit ebca6153

.
Signed-off-by: Jeremy Berchtold <jberchtold@nvidia.com>

58c3ac80

Bump cuDNN FE to 1.14.0 (#2072) · d2615d1c

vcherepanov-nv authored Aug 26, 2025



* Bump cuDNN FE to 1.14.0
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Change submodule hash
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Pick up a cuDNN FE fix
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* New model configs in tests
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

* Exclude cuDNN backend for some configs
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

---------
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

d2615d1c

Revert "[Common] PDL for Quantization Kernels" (#2114) · 4572dbef

jberchtold-nvidia authored Aug 26, 2025

Revert "[Common] PDL for Quantization Kernels (#2001)"

This reverts commit bfab8c67

.
Signed-off-by: Jeremy Berchtold <jberchtold@nvidia.com>

4572dbef

[JAX] Add Shardy warning in GEMM custom call (#2101) · c638ac7e

Phuong Nguyen authored Aug 25, 2025



* added shardy warning
Signed-off-by: Phuong Nguyen <phuonguyen@nvidia.com>


---------
Signed-off-by: Phuong Nguyen <phuonguyen@nvidia.com>

c638ac7e

[PyTorch] Debug Mcore wgrad fusion with te.ops (#2097) · e94041a5

Tim Moon authored Aug 22, 2025



* Return dummy wgrad tensors when requested by Mcore
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Apply suggestions from code review
Co-authored-by: Jan Bielak <janekb04@icloud.com>
Signed-off-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>

---------
Signed-off-by: Tim Moon <tmoon@nvidia.com>
Signed-off-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>
Co-authored-by: Jan Bielak <janekb04@icloud.com>

e94041a5

[ TE-JAX ] Expose cp_strategy argument to DPA api (#2090) · 0168c268

Md Fahim Faysal Khan authored Aug 21, 2025



* added cp strategy arg to DPA api
Signed-off-by: Md Fahim Faysal Khan <mdfahimfaysa@nvidia.com>

* converted DPA cp_strategy to string
Signed-off-by: Md Fahim Faysal Khan <mdfahimfaysa@nvidia.com>

---------
Signed-off-by: Md Fahim Faysal Khan <mdfahimfaysa@nvidia.com>

0168c268

Fix incorrect version checks for atomic GEMM (#2095) · 3a4136b6

Tim Moon authored Aug 20, 2025



* Fix incorrect version checks for atomic GEMM
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Fix typo
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci



---------
Signed-off-by: Tim Moon <tmoon@nvidia.com>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

3a4136b6

[PyTorch] Avoid garbage collection when capturing a CUDA Graph (#2092) · 9f065fa2
Tim Moon authored Aug 20, 2025
```
Avoid garbage collection when capturing a CUDA Graph
Signed-off-by: Tim Moon <tmoon@nvidia.com>
```
9f065fa2

[JAX] Error checking for mesh resource and update GemmPrimitive to use... · 34150d1a

jberchtold-nvidia authored Aug 20, 2025


[JAX] Error checking for mesh resource and update GemmPrimitive to use global_mesh_resource().fsdp_resource (#2088)

* Enforce global MeshResource is set
Signed-off-by: Jeremy Berchtold <jberchtold@nvidia.com>

* Use global_mesh_resource().fsdp_resource in gemm primitive
Signed-off-by: Jeremy Berchtold <jberchtold@nvidia.com>

* Update tests
Signed-off-by: Jeremy Berchtold <jberchtold@nvidia.com>

* Update gemm.py
Signed-off-by: Jeremy Berchtold <jberchtold@nvidia.com>

* Update test_layer.py
Signed-off-by: Jeremy Berchtold <jberchtold@nvidia.com>

---------
Signed-off-by: Jeremy Berchtold <jberchtold@nvidia.com>

34150d1a

[DCU] fix · 4927d10e
yuguo authored Aug 26, 2025

4927d10e
[DCU] fix · 2e870ed9
yuguo authored Aug 26, 2025

2e870ed9

25 Aug, 2025 1 commit
- [DCU] fix moe tensorwise int8 · 059d92e2
  yuguo authored Aug 25, 2025
  
  059d92e2
23 Aug, 2025 2 commits
- [DCU] tensorwise int8 gemm surpport bias · 11864d3d
  yuguo authored Aug 23, 2025
  
  11864d3d
- [DCU] fix tensorwise int8 moe bugs · 32edae18
  yuguo authored Aug 23, 2025
  
  32edae18
21 Aug, 2025 2 commits
- fix · 0cf10d1c
  yuguo authored Aug 21, 2025
  
  0cf10d1c
- [DCU] tensorwise int8 train opt · 7a923605
  yuguo authored Aug 21, 2025
  
  7a923605
19 Aug, 2025 1 commit
- add swap env · d19a5a44
  evt_fugx1 authored Aug 19, 2025
  
  d19a5a44
18 Aug, 2025 3 commits

[JAX] Fix for TE GEMM - Always AllGather RHS non-contracting dims with FSDP axis (#2075) · 3fc1e4bf
Phuong Nguyen authored Aug 18, 2025
```
* fix fsdp
Signed-off-by: Phuong Nguyen <phuonguyen@nvidia.com>

---------
Signed-off-by: Phuong Nguyen <phuonguyen@nvidia.com>
```
3fc1e4bf

[PyTorch] Check if the given recipe is supported in `fp8_autocast` (#2073) · 0e3e270f

Xin Yao authored Aug 19, 2025



* check if the given recipe is supported in fp8_autocast
Signed-off-by: Xin Yao <xiny@nvidia.com>

* resolve comments
Signed-off-by: Xin Yao <xiny@nvidia.com>

* check only when enabled
Signed-off-by: Xin Yao <xiny@nvidia.com>

---------
Signed-off-by: Xin Yao <xiny@nvidia.com>
Co-authored-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>

0e3e270f

[JAX] Fix Flax variable creation when quantizers are created directly from a recipe (#2079) · 757fd1cf
jberchtold-nvidia authored Aug 18, 2025
```
Fix flax variables when creating quantizers directly from a recipe
Signed-off-by: Jeremy Berchtold <jberchtold@nvidia.com>
```
757fd1cf