Skip to content

[None][feat] Update CuTeDSL MegaMoE kernels #16190

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
Barry-Delaney wants to merge 10 commits into
base: main
Choose a base branch
from
+7,727 −3,816
Draft

[None][feat] Update CuTeDSL MegaMoE kernels #16190

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
10 commits
Select commit Hold shift + click to select a range
a7e71fb
[TRTLLM-12950][feat] MegaMoE-CuteDSL fused-MoE backend (DSv4 DEP8, NV…
Barry-Delaney Jul 3, 2026
b7eed33
[TRTLLM-12950][test] reproducible MegaMoE-CuteDSL kernel microbench
Barry-Delaney Jul 3, 2026
6e8d40d
[TRTLLM-12950][feat] MegaMoE-CuteDSL: stream source weights through t…
Barry-Delaney Jul 8, 2026
faf2f1c
[TRTLLM-12950][fix] MegaMoE-CuteDSL: address review round 2 (integrat…
Barry-Delaney Jul 9, 2026
2e47514
[TRTLLM-12950][fix] MegaMoE-CuteDSL vendored kernel: Int64 SF indexin…
Barry-Delaney Jul 9, 2026
28813b8
[TRTLLM-12950][fix] MegaMoE-CuteDSL: address review round 3 (integrat…
Barry-Delaney Jul 9, 2026
619b273
[TRTLLM-12950][fix] MegaMoE-CuteDSL vendored kernel: restore IKET Not…
Barry-Delaney Jul 9, 2026
6d3af84
[TRTLLM-12950][fix] MegaMoE-CuteDSL: address review round 4 (reload p…
Barry-Delaney Jul 9, 2026
0cf0566
[TRTLLM-12950][fix] MegaMoE-CuteDSL: adapt streamed-load hooks to mai…
Barry-Delaney Jul 9, 2026
cbdc947
[TRTLLM-12950][fix] MegaMoE-CuteDSL: address review round 5 (reload o…
Barry-Delaney Jul 9, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 3 additions & 1 deletion .pre-commit-config.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -304,6 +304,7 @@ common-files: &common_files |
tensorrt_llm/_torch/cute_dsl_kernels/blackwell/utils.py |
tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/__init__.py |
tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/blocked_scale.py |
tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/common_moe_utils.py |
tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/contract.py |
tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/custom_ext.py |
tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/dynamic_mainloop.py |
Expand Down Expand Up @@ -1678,6 +1679,7 @@ legacy-files: &legacy_files |
tensorrt_llm/_torch/cute_dsl_kernels/blackwell/utils.py |
tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/__init__.py |
tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/blocked_scale.py |
tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/common_moe_utils.py |
tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/contract.py |
tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/custom_ext.py |
tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/dynamic_mainloop.py |
Expand Down Expand Up @@ -2835,7 +2837,7 @@ repos:
additional_dependencies:
- tomli
# add ignore words list
args: ["-L", "Mor,ans,thirdparty,subtiles,PARD,pard,indx,therefrom", "--skip", "ATTRIBUTIONS-*.md,*.svg", "--skip", "security_scanning/*", "--skip", "tensorrt_llm/_torch/visual_gen/jit_kernels/*"]
args: ["-L", "Mor,ans,thirdparty,subtiles,PARD,pard,indx,therefrom", "--skip", "ATTRIBUTIONS-*.md,*.svg", "--skip", "security_scanning/*", "--skip", "tensorrt_llm/_torch/visual_gen/jit_kernels/*", "--skip", "tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/*"]
exclude: 'scripts/attribution/data/cas/.*$'
- repo: https://github.com/astral-sh/ruff-pre-commit
rev: v0.9.4
Expand Down
1 change: 1 addition & 0 deletions legacy-files.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -296,6 +296,7 @@ tensorrt_llm/_torch/cute_dsl_kernels/blackwell/dense_blockscaled_gemm_persistent
tensorrt_llm/_torch/cute_dsl_kernels/blackwell/utils.py
tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/__init__.py
tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/blocked_scale.py
tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/common_moe_utils.py
tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/contract.py
tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/custom_ext.py
tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/dynamic_mainloop.py
Expand Down
1 change: 1 addition & 0 deletions pyproject.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -353,6 +353,7 @@ exclude = [
"tensorrt_llm/_torch/cute_dsl_kernels/blackwell/utils.py",
"tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/__init__.py",
"tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/blocked_scale.py",
"tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/common_moe_utils.py",
"tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/contract.py",
"tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/custom_ext.py",
"tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/dynamic_mainloop.py",
Expand Down
3 changes: 3 additions & 0 deletions ruff-legacy-baseline.json
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -524,6 +524,9 @@
"tensorrt_llm/_torch/cute_dsl_kernels/blackwell/dense_blockscaled_gemm_persistent.py": {
"E741": 3
},
"tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/kernel_fc12.py": {
"E731": 1
},
"tensorrt_llm/_torch/distributed/ops.py": {
"E712": 1,
"E731": 1
Expand Down
1 change: 1 addition & 0 deletions ruff-legacy.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -313,6 +313,7 @@ include = [
"tensorrt_llm/_torch/cute_dsl_kernels/blackwell/utils.py",
"tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/__init__.py",
"tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/blocked_scale.py",
"tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/common_moe_utils.py",
"tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/contract.py",
"tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/custom_ext.py",
"tensorrt_llm/_torch/cute_dsl_kernels/mega_moe_nvfp4/dynamic_mainloop.py",
Expand Down
Loading
Loading