Lixun Zhang 2e217c5a5c [Backend] Refactor sharedToDotOperandMFMA lowering (#439) ...

* Remove unnecessary xor computations for k-major swizzled tensors

* Support mfma16 and mfma4 in the fast path

* Choose warpsPerCTA according to nonKDim

* Set maxPhase=4 for mfma4

* Fix tests

For now, we do not disable swizzling for k-major tensors

* Remove fastPathComputeOffsetsTy1

* Enable k-major + disabled swizzling in the normal path

2024-01-12 12:50:18 -06:00

..

Analysis

enable layout conversion from mfma to dot_op for mfma16. (#453)

2024-01-10 22:31:59 -06:00

Conversion

[Backend] Refactor sharedToDotOperandMFMA lowering (#439)

2024-01-12 12:50:18 -06:00

Dialect

[Backend] Refactor sharedToDotOperandMFMA lowering (#439)

2024-01-12 12:50:18 -06:00

Target

Update hip_backend to use libhsa-runtime for arch info, (#411)

2023-12-21 15:40:57 +00:00

CMakeLists.txt

[BACKEND] Remove HopperHelpers.c and replace with inline ptx and LLVM codegen (#2047)

2023-08-10 15:52:37 -07:00