Momchil Velikov 73b85f87e4

[MLIR][AArch64] Lower vector.contract to SVE FEAT_BF16 operations (#147052) ...

This patch adds lowering of Bfloat16 widening matrix multiply and
accumulate `vector.contract`, by parametrising and refactoring the
pattern for 8-bit integers.

2025-07-22 10:09:28 +01:00

..

ArmSVE

[MLIR][AArch64] Lower vector.contract to SVE FEAT_BF16 operations (#147052)

2025-07-22 10:09:28 +01:00

X86

…