ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-04-19 22:39:03 +00:00
Code Issues Packages Projects Releases Wiki Activity

[CK][Examples] Extending support for rdna3/4 part 2:

Browse Source 
-example_batched_gemm_xdl_int8
-example_batched_gemm_xdl_fp8_rowwise_v3
-example_batched_gemm_xdl_fp32
-example_batched_gemm_xdl_bf16
-example_batched_gemm_xdl_bf16_v3
-example_batched_gemm_xdl_fp16
-example_splitk_gemm_bias_e_permute_xdl_fp32
*fixing return value to return 0 as success in above examples.

Fixing cmdline parameters in:
-example_sparse_embedding3_forward_layernorm
-example_elementwise_binary_4D_fp16
-elementwise_scale_permute_amax_2D_fp16_fp8

Signed-off-by: Michal Kulikowski <Michal.Kulikowski@amd.com>
This commit is contained in:
Michal Kulikowski
2025-10-01 16:04:25 +02:00
committed by Michał Kulikowski
parent 1d4db30af9
commit 7259b9c4db
16 changed files with 85 additions and 73 deletions
Download Patch File Download Diff File Expand all files Collapse all files

23
example/36_sparse_embedding/sparse_embedding3_forward_layernorm.cpp
View File

@@ -1,5 +1,5 @@
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
// Copyright (c) 2018-2025, Advanced Micro Devices, Inc. All rights reserved.
#include <iostream>
#include <numeric>
@@ -77,17 +77,21 @@ int main(int argc, char* argv[])
{
// Use default value
}
else if(argc == 4)
else if(argc == 5)
{
num_rows = atoi(argv[1]);
dim_mask = strtol(argv[2], nullptr, 0);
index_length = atoi(argv[3]);
time_kernel = std::stoi(argv[1]);
num_rows = std::stoi(argv[2]);
dim_mask = strtol(argv[3], nullptr, 0);
index_length = std::stoi(argv[4]);
}
else
{
std::cout << "Usage of " << argv[0] << std::endl;
std::cout << "Arg1-3: num_rows dim_mask index_length" << std::endl;
std::cout << "arg1: time kernel (0=no, 1=yes)" << std::endl;
std::cout << "arg2-4: num_rows dim_mask index_length" << std::endl;
return 1;
}
ck::static_for<0, dims.Size(), 1>{}([&](auto I) {
if(dim_mask & (1 << I.value))
{
@@ -160,11 +164,10 @@ int main(int argc, char* argv[])
<< std::endl
<< std::flush;
bool is_supported = device_instance.IsSupportedArgument(argument_ptr.get());
if(!is_supported)
if(!device_instance.IsSupportedArgument(argument_ptr.get()))
{
std::cout << "Runtime parameters are not supported" << std::endl;
std::cerr << device_instance.GetTypeString() << " does not support this problem"
<< std::endl;
return;
}