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

fix: fix bug in print tile window when printing bf8/fp8 tiles (#3120)

Browse Source 
* fix: fix bug in print tile window when printing bf8/fp8 tiles

* test(print_tile_window_range): add unit tests to maintain function integrity

* fix: fp8 numerical mismatch error on gfx950 by adding DCK_TILE_USE_OCP_FP8

[ROCm/composable_kernel commit: 45be741586]
This commit is contained in:
Aviral Goel
2025-11-01 15:28:07 -04:00
committed by GitHub
parent 515fb27488
commit 69f7ade10b
3 changed files with 230 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
include/ck_tile/core/tensor/tile_window.hpp
View File

@@ -1076,7 +1076,7 @@ struct tile_window_with_static_lengths
using ThreadBuf = thread_buffer<DataType, 2>;
auto buf = tensor_view.template get_vectorized_elements<ThreadBuf>(coord, 0);
auto value = buf.at(number<0>{}); // Extract first element from thread buffer
printf(" %s[%d,%d] = %f", label, i, j, static_cast<float>(value));
printf(" %s[%d,%d] = %f", label, i, j, type_convert<float>(value));
}
printf("\n");
}