ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-19 12:30:16 +00:00
Code Issues Packages Projects Releases Wiki Activity

Rangify constructor of HostTensorDescriptor & Tensor<> (#445)

Browse Source 
* Rangify STL algorithms

This commit adapts rangified std::copy(), std::fill() & std::transform()

* Rangify check_err()

By rangifying check_err(), we can not only compare values between
std::vector<>s, but also compare any ranges which have same value
type.

* Allow constructing Tensor<> like a HostTensorDescriptor

* Simplify Tensor<> object construction logics

* Remove more unnecessary 'HostTensorDescriptor' objects

* Re-format example code

* Re-write more HostTensorDescriptor ctor call

[ROCm/composable_kernel commit: 4a2a56c22f]
This commit is contained in:
Po Yen Chen
2022-11-12 01:36:01 +08:00
committed by GitHub
parent a4d6101e99
commit f2dd2e5b09
103 changed files with 657 additions and 649 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
example/24_batched_gemm/run_batched_gemm_example.inc
View File

@@ -55,15 +55,15 @@ bool run_batched_gemm(const ProblemSize& problem_size, const ExecutionConfig& co
std::size_t stride,
std::size_t batch_stride,
auto layout) {
using namespace ck::literals;
if(std::is_same<decltype(layout), ck::tensor_layout::gemm::RowMajor>::value)
{
return HostTensorDescriptor(std::vector<std::size_t>({batch_count_, row, col}),
std::vector<std::size_t>({batch_stride, stride, 1}));
return HostTensorDescriptor({batch_count_, row, col}, {batch_stride, stride, 1_uz});
}
else
{
return HostTensorDescriptor(std::vector<std::size_t>({batch_count_, row, col}),
std::vector<std::size_t>({batch_stride, 1, stride}));
return HostTensorDescriptor({batch_count_, row, col}, {batch_stride, 1_uz, stride});
}
};
@@ -174,11 +174,11 @@ bool run_batched_gemm(const ProblemSize& problem_size, const ExecutionConfig& co
#ifdef BUILD_INT4_EXAMPLE
const Tensor<EDataType> e_device_result_converted(e_g_m_n_device_result);
pass &= ck::utils::check_err(e_device_result_converted.mData, e_g_m_n_host_result.mData);
pass &= ck::utils::check_err(e_device_result_converted, e_g_m_n_host_result);
#else
pass = ck::utils::check_err(
e_g_m_n_device_result.mData, e_g_m_n_host_result.mData, "Error: Incorrect results c");
e_g_m_n_device_result, e_g_m_n_host_result, "Error: Incorrect results c");
#endif
}