Update to gpu_timer for rotating_buffer (#2524)

* update gpu_timer for rotating buffer as hipblasLt's implementation

* timing fix

* Updating gpu timer for old ck as well

* Revert "Updating gpu timer for old ck as well"

This reverts commit 958cd1bc99.

* code clean up with runtime argument; function rename

* code cleanup

* general timer fixes

* bug fix

* clang formatted

* addressing reveiew comments

* clang formatted

* Addressing review comments

* CI fix

---------

Co-authored-by: Po Yen Chen <PoYen.Chen@amd.com>
This commit is contained in:
Khushbu Agarwal
2025-07-29 15:21:05 -07:00
committed by GitHub
parent b80099cc5f
commit 61e21f5567
13 changed files with 182 additions and 78 deletions

View File

@@ -147,18 +147,14 @@ float gemm(const ck_tile::GemmHostArgs& args, const ck_tile::stream_config& s)
if(s.flush_cache_)
{
std::cout << "Flushing cache..." << std::endl;
static constexpr ck_tile::index_t APackedSize =
std::is_same_v<BDataType, ck_tile::pk_int4_t> ? 2 : 1;
static constexpr ck_tile::index_t BPackedSize =
std::is_same_v<BDataType, ck_tile::pk_int4_t> ? 2 : 1;
ck_tile::HostTensor<ADataType> a_m(ck_tile::host_tensor_descriptor(
args.M, args.K, args.stride_A, is_row_major(ALayout{})));
ck_tile::HostTensor<BDataType> b_n(ck_tile::host_tensor_descriptor(
args.K, args.N, args.stride_B, is_row_major(BLayout{})));
auto size_a_buffer = a_m.get_element_space_size_in_bytes() / APackedSize;
auto size_b_buffer = b_n.get_element_space_size_in_bytes() / BPackedSize;
auto size_a_buffer = a_m.get_element_space_size_in_bytes();
auto size_b_buffer = b_n.get_element_space_size_in_bytes();
ck_tile::RotatingMemWrapper<ADataType, BDataType> rotating_mem(
kargs.as_ptr[0], kargs.bs_ptr[0], s.rotating_count_, size_a_buffer, size_b_buffer);
@@ -174,7 +170,7 @@ float gemm(const ck_tile::GemmHostArgs& args, const ck_tile::stream_config& s)
hipGetErrorString(hipMemsetAsync(
args.e_ptr, 0, args.M * args.N * sizeof(CDataType), s.stream_id_));
};
ave_time = ck_tile::launch_kernel_preprocess(
ave_time = ck_tile::launch_kernel_time_mask(
s,
run_flush_cache,
ck_tile::make_kernel<blocks.x, GemmConfig::kBlockPerCu>(