Repository navigation
Conversation
Signed-off-by: Kaining Zhong <kainingz@nvidia.com>
for more information, see https://pre-commit.ci
Signed-off-by: Kaining Zhong <kainingz@nvidia.com>
for more information, see https://pre-commit.ci
Signed-off-by: Kaining Zhong <kainingz@nvidia.com>
Signed-off-by: Kaining Zhong <kainingz@nvidia.com>
Signed-off-by: Kaining Zhong <kainingz@nvidia.com>
Signed-off-by: Kaining Zhong <kainingz@nvidia.com>
Signed-off-by: Kaining Zhong <kainingz@nvidia.com>
Signed-off-by: Kaining Zhong <kainingz@nvidia.com>
Signed-off-by: Kaining Zhong <kainingz@nvidia.com>
for more information, see https://pre-commit.ci
Signed-off-by: Kaining Zhong <kainingz@nvidia.com>
The more_cutedsl_mxfp8_specialized base was squash-merged as NVIDIA#3642, so CuTeDSL conflicts were resolved by 3-way merging against 44d2c22, the last base-branch commit merged here. CuTeDSL/__init__.py: register both get_mxfp8_dequantization_function and NVIDIA#3560's get_mxfp8_requantization_function. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Signed-off-by: Kaining Zhong <kainingz@nvidia.com>
 [Medium impact] The change reaches MXFP8 dequantization in the shared cast path. Fix the missing reporting function calls before merging; they block builds with CuTeDSL support. Findings
SummaryAdds CuTeDSL kernels for MXFP8 dequantization.
Builds with CuTeDSL support are blocked by calls to a missing reporting function. Runtime tests were not run. Diagram%%{init: {'theme': 'neutral'}}%%
flowchart TD
A[MXFP8 dequantize] --> B{CuTeDSL enabled and supported?}
B -->|No| C[CUDA kernel]
B -->|Yes| D[Load or compile registered kernel]
D --> E{Kernel available?}
E -->|No| C
E -->|Yes| F[Choose rowwise data if present]
F --> G[Load tiles and apply scales]
G --> H[Write higher-precision output]
Reviews (1) · Last reviewed commit: "Merge branch 'main' into cutedsl_mxfp8_d..." · Reviewed by Greptile |
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
| // Signature mirrors mxfp8::dequantize. Returns false to fall back to the CUDA kernel. | ||
| inline bool mxfp8_dequantize_cutedsl(const Tensor &input, Tensor *output, cudaStream_t stream) { | ||
| if (!tvm_ffi_bridge::TVMFFICentral::getInstance().get_cutedsl_backend_enabled()) { | ||
| maybe_warn_cutedsl_not_chosen("the CuTeDSL backend is disabled."); |
There was a problem hiding this comment.
maybe_warn_cutedsl_not_chosen is not declared or defined anywhere in the repository. Builds with NVTE_WITH_CUTEDSL include this header and fail to compile, even when the backend is disabled at runtime. Replace all three calls here and at lines 105 and 114 with TVMFFICentral::getInstance().report_not_chosen(...), the existing reporting method.
Description
Add MXFP8 dequantization CuTeDSL kernels
Fixes # (issue)
Type of change
Changes
Please list the changes introduced in this PR:
Checklist: