Adds platform-specific warp mask to frag_layout_swizzle.cuh#7
Merged
diptorupd merged 1 commit intoamd-integrationfrom Oct 2, 2025
Merged
Conversation
There was a problem hiding this comment.
Pull Request Overview
This PR adds platform-specific warp mask constants to support both CUDA (32-thread warps) and HIP/CDNA3 (64-thread warps) architectures. The change replaces hardcoded 32-bit masks with a platform-conditional WARP_FULL_MASK constant.
- Introduces platform-specific
WARP_FULL_MASKconstant (32-bit for CUDA, 64-bit for HIP) - Replaces hardcoded
0xffffffffmasks in__shfl_xor_synccalls with the new constant - Enables compatibility with CDNA3's 64-thread warps while maintaining CUDA support
Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.
libflashinfer/include/flashinfer/attention/generic/frag_layout_swizzle.cuh
Show resolved
Hide resolved
demandal25
approved these changes
Oct 2, 2025
diptorupd
pushed a commit
that referenced
this pull request
Dec 5, 2025
In this PR I remove the `libtorch` dependency and removed
`test_page.cpp`. `test_page.cpp` is the only unit test that uses
libtorch. However, we also have a pytest for testing page. We will use
that for validation.
Removing the libtorch dependency will help us speed docker builds and
remove additional dependencies.
```Test project /root/flashinfer/libflashinfer/tests/hip/build
Start 1: MathTest
1/8 Test #1: MathTest ............................ Passed 0.31 sec
Start 2: PosEncTest
2/8 Test #2: PosEncTest .......................... Passed 0.31 sec
Start 3: CascadeTest
3/8 Test #3: CascadeTest ......................... Passed 1369.12 sec
Start 4: SingleDecodeTest
4/8 Test #4: SingleDecodeTest .................... Passed 7726.35 sec
Start 5: BatchDecodeTest
5/8 Test #5: BatchDecodeTest ..................... Passed 811.61 sec
Start 6: test_mfma_fp32_16x16x16fp16
6/8 Test #6: test_mfma_fp32_16x16x16fp16 ......... Passed 0.30 sec
Start 7: test_transpose_4x4_half_registers
7/8 Test #7: test_transpose_4x4_half_registers ... Passed 0.28 sec
Start 8: test_rowsum
8/8 Test #8: test_rowsum ......................... Passed 0.27 sec
100% tests passed, 0 tests failed out of 8
```
diptorupd
added a commit
that referenced
this pull request
Dec 5, 2025
The mask values for the __shfl_xor_sync used in the original CUDA version of the frag_layout_swizzle.cuh was designed for 32-thread warps. The design made the header incompatible with CDNA3 that has 64-thread warps. The PR adds a platform-specific WARP_FULL_MASK constant to support both 32 thread and 64 thread warps.
zhenhantech
pushed a commit
to zhenhantech/flashinfer
that referenced
this pull request
Jan 9, 2026
In this PR I remove the `libtorch` dependency and removed
`test_page.cpp`. `test_page.cpp` is the only unit test that uses
libtorch. However, we also have a pytest for testing page. We will use
that for validation.
Removing the libtorch dependency will help us speed docker builds and
remove additional dependencies.
```Test project /root/flashinfer/libflashinfer/tests/hip/build
Start 1: MathTest
1/8 Test ROCm#1: MathTest ............................ Passed 0.31 sec
Start 2: PosEncTest
2/8 Test ROCm#2: PosEncTest .......................... Passed 0.31 sec
Start 3: CascadeTest
3/8 Test ROCm#3: CascadeTest ......................... Passed 1369.12 sec
Start 4: SingleDecodeTest
4/8 Test ROCm#4: SingleDecodeTest .................... Passed 7726.35 sec
Start 5: BatchDecodeTest
5/8 Test ROCm#5: BatchDecodeTest ..................... Passed 811.61 sec
Start 6: test_mfma_fp32_16x16x16fp16
6/8 Test ROCm#6: test_mfma_fp32_16x16x16fp16 ......... Passed 0.30 sec
Start 7: test_transpose_4x4_half_registers
7/8 Test ROCm#7: test_transpose_4x4_half_registers ... Passed 0.28 sec
Start 8: test_rowsum
8/8 Test ROCm#8: test_rowsum ......................... Passed 0.27 sec
100% tests passed, 0 tests failed out of 8
```
zhenhantech
pushed a commit
to zhenhantech/flashinfer
that referenced
this pull request
Jan 9, 2026
The mask values for the __shfl_xor_sync used in the original CUDA version of the frag_layout_swizzle.cuh was designed for 32-thread warps. The design made the header incompatible with CDNA3 that has 64-thread warps. The PR adds a platform-specific WARP_FULL_MASK constant to support both 32 thread and 64 thread warps.
diptorupd
pushed a commit
that referenced
this pull request
Feb 2, 2026
In this PR I remove the `libtorch` dependency and removed
`test_page.cpp`. `test_page.cpp` is the only unit test that uses
libtorch. However, we also have a pytest for testing page. We will use
that for validation.
Removing the libtorch dependency will help us speed docker builds and
remove additional dependencies.
```Test project /root/flashinfer/libflashinfer/tests/hip/build
Start 1: MathTest
1/8 Test #1: MathTest ............................ Passed 0.31 sec
Start 2: PosEncTest
2/8 Test #2: PosEncTest .......................... Passed 0.31 sec
Start 3: CascadeTest
3/8 Test #3: CascadeTest ......................... Passed 1369.12 sec
Start 4: SingleDecodeTest
4/8 Test #4: SingleDecodeTest .................... Passed 7726.35 sec
Start 5: BatchDecodeTest
5/8 Test #5: BatchDecodeTest ..................... Passed 811.61 sec
Start 6: test_mfma_fp32_16x16x16fp16
6/8 Test #6: test_mfma_fp32_16x16x16fp16 ......... Passed 0.30 sec
Start 7: test_transpose_4x4_half_registers
7/8 Test #7: test_transpose_4x4_half_registers ... Passed 0.28 sec
Start 8: test_rowsum
8/8 Test #8: test_rowsum ......................... Passed 0.27 sec
100% tests passed, 0 tests failed out of 8
```
diptorupd
added a commit
that referenced
this pull request
Feb 2, 2026
The mask values for the __shfl_xor_sync used in the original CUDA version of the frag_layout_swizzle.cuh was designed for 32-thread warps. The design made the header incompatible with CDNA3 that has 64-thread warps. The PR adds a platform-specific WARP_FULL_MASK constant to support both 32 thread and 64 thread warps.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The mask values for the
__shfl_xor_syncused in the original CUDA version of the frag_layout_swizzle.cuh was designed for 32-thread warps. The design made the header incompatible with CDNA3 that has 64-thread warps. The PR adds a platform-specificWARP_FULL_MASKconstant to support both 32 thread and 64 thread warps.