ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-14 02:02:46 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge commit '1eb26460aa621028c5d5a8a20cf593ed8a3a3cc5' into develop

Browse Source
This commit is contained in:
assistant-librarian[bot]
2025-11-19 15:13:07 +00:00
parent 055f643f4f
commit 2e77585ae1
2 changed files with 10 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
example/ck_tile/36_pooling/pool3d.cpp
View File

@@ -31,8 +31,8 @@ auto create_args(int argc, char* argv[])
.insert("RightPy", "1", "right padding h")
.insert("RightPx", "1", "right padding w")
.insert("v", "1", "cpu validation or not")
.insert("warmup", "0", "cold iter")
.insert("repeat", "1", "hot iter");
.insert("warmup", "20", "cold iter")
.insert("repeat", "100", "hot iter");
bool result = arg_parser.parse(argc, argv);
return std::make_tuple(result, arg_parser);
@@ -120,10 +120,10 @@ bool run(const ck_tile::ArgParser& arg_parser)
in_buf.ToDevice(in.data());
using ReduceOp = ck_tile::ReduceOp::Max;
using BlockWarps = ck_tile::sequence<4, 1>;
using BlockTile = ck_tile::sequence<128, 128>;
using WarpTile = ck_tile::sequence<32, 128>;
using ThreadTile = ck_tile::sequence<8, 8>;
using BlockWarps = ck_tile::sequence<1, 1>;
using BlockTile = ck_tile::sequence<128, 1>;
using WarpTile = ck_tile::sequence<128, 1>;
using ThreadTile = ck_tile::sequence<2, 1>;
using Shape = ck_tile::PoolShape<BlockWarps, BlockTile, WarpTile, ThreadTile>;
using Problem = ck_tile::PoolProblem<InDataType,

8
test/ck_tile/pooling/test_pooling.cpp
View File

@@ -295,10 +295,10 @@ class TestCkTilePooling : public ::testing::Test
}
};
using Shape1_BlockWarps = ck_tile::sequence<4, 1>;
using Shape1_BlockTile = ck_tile::sequence<128, 128>;
using Shape1_WarpTile = ck_tile::sequence<32, 128>;
using Shape1_ThreadTile = ck_tile::sequence<8, 8>;
using Shape1_BlockWarps = ck_tile::sequence<1, 1>;
using Shape1_BlockTile = ck_tile::sequence<128, 1>;
using Shape1_WarpTile = ck_tile::sequence<128, 1>;
using Shape1_ThreadTile = ck_tile::sequence<2, 1>;
// Cross-warp configuration
using Shape2_BlockWarps = ck_tile::sequence<2, 2>;