leavelet/cutlass
1
0
Fork 0
Code Issues Pull Requests Actions 26 Packages Projects Releases Wiki Activity

CUTLASS 3.3.0 (#1167)

Browse Source 
* Release 3.3.0

Adds support for mixed precision GEMMs On Hopper and Ampere
Adds support for < 16B aligned GEMMs on Hopper
Enhancements to EVT
Enhancements to Python interface
Enhancements to Sub-byte type handling in CuTe
Several other bug-fixes and performance improvements.

* minor doc update
This commit is contained in:
Pradeep Ramani
2023-11-02 11:09:05 -04:00
committed by GitHub
parent 922fb5108b
commit c008b4aea8
263 changed files with 16214 additions and 5008 deletions
Download Patch File Download Diff File Expand all files Collapse all files
test/python/cutlass/gemm/gemm_f16_sm90.py
+4
View File
@@ -135,6 +135,10 @@ add_test_simt(layouts=LayoutCombination.NTN, element_output=cutlass.DataType.f16
add_test_simt(layouts=LayoutCombination.TTN, element_output=cutlass.DataType.f16, element_accumulator=cutlass.DataType.f32, threadblock_shape=[ 64, 64, 8])
add_test_simt(layouts=LayoutCombination.NNT, element_output=cutlass.DataType.f16, element_accumulator=cutlass.DataType.f16, threadblock_shape=[128, 128, 8])
# Tests with void-C kernels
add_test_cluster_shape(layouts=LayoutCombination.NNT, alignments=[8, 8, 8], element_output=cutlass.DataType.f16,
element_accumulator=cutlass.DataType.f32, threadblock_shape=[128, 128, 32], stages=None,
cluster_shape=[2, 1, 1], element_C=cutlass.DataType.void)
if __name__ == '__main__':
unittest.main()