Po Yen Chen f7cc8c3b03 Update tuning parameter & compilation options of DeviceGemmXdl<> instance (layout=TT) (#819) ...

* Enable pipeline v2 opt for layout=TT instance

* Use better thread mapping for reading A tile

* Conditionally enable pipeline v2 opt

* Allow enabling only fp16 gemm instances in profiler

* Fix formatting error

* Fix compilation error if we enable fp32 in profiler

2023-08-02 10:32:22 -05:00

..

tensor_operation_instance/gpu

Update tuning parameter & compilation options of DeviceGemmXdl<> instance (layout=TT) (#819)

2023-08-02 10:32:22 -05:00

utility

initial stream-k implementation with example (#699)

2023-07-26 14:18:15 -05:00