profiler/src/profile_grouped_conv_fwd.cpp · fd0d093e78c18197a4f1b7dafdbc1e2438d28317 · gaoqiong / composable_kernel_ROCM

Rostyslav Geyyer authored Mar 21, 2024

* Add bf8 conv fwd instances

* Add example

* Add profiler mode

* Add client example

* Fix copyright headers

* Format

fd0d093e

profile_grouped_conv_fwd.cpp 9.75 KB