diff options
| author | Georgios Pinitas <georgios.pinitas@arm.com> | 2021-07-16 16:16:43 +0100 |
|---|---|---|
| committer | Georgios Pinitas <georgios.pinitas@arm.com> | 2021-07-22 02:25:50 +0000 |
| commit | 4ee8b1599dbaf7634d25607fa5ac96ba3dc6b0f2 (patch) | |
| tree | 2f8362d33cdad4212f4b96995681c68184c759e1 /src/core/NEON/kernels/arm_gemm/transforms/list.hpp | |
| parent | 59fd7a722e5bc7e85309d6200bc37a772721a719 (diff) | |
| download | ComputeLibrary-4ee8b1599dbaf7634d25607fa5ac96ba3dc6b0f2.tar.gz | |
Update GEMM assembly kernels
- Introduce Fp32 kernels with internal calculations in Bfloat16 when
fast_mode is enabled
- Improve kernel selection heuristics
Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com>
Change-Id: I68a9e7e862b6fd2721b46e0d7cc791091c4ab279
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5965
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/core/NEON/kernels/arm_gemm/transforms/list.hpp')
| -rw-r--r-- | src/core/NEON/kernels/arm_gemm/transforms/list.hpp | 29 |
1 files changed, 25 insertions, 4 deletions
diff --git a/src/core/NEON/kernels/arm_gemm/transforms/list.hpp b/src/core/NEON/kernels/arm_gemm/transforms/list.hpp index e092c729ba..adbaa6cf2f 100644 --- a/src/core/NEON/kernels/arm_gemm/transforms/list.hpp +++ b/src/core/NEON/kernels/arm_gemm/transforms/list.hpp @@ -22,7 +22,28 @@ * SOFTWARE. */ #include "a32_transpose_interleave_8way_32bit.hpp" -#include "a64_transpose_interleave_12way_16bit.hpp" -#include "a64_transpose_interleave_12way_half_to_float.hpp" -#include "a64_transpose_interleave_24way_16bit.hpp" -#include "a64_transpose_interleave_8way_32bit.hpp" +#include "a64_transpose_interleave_12_1x4.hpp" +#include "a64_transpose_interleave_12_1x8.hpp" +#include "a64_transpose_interleave_12_2x2.hpp" +#include "a64_transpose_interleave_12_2x4_fp32bf16.hpp" +#include "a64_transpose_interleave_12_2x4.hpp" +#include "a64_transpose_interleave_128.hpp" +#include "a64_transpose_interleave_12_s8s16.hpp" +#include "a64_transpose_interleave_12_u8u16.hpp" +#include "a64_transpose_interleave_16_1x4.hpp" +#include "a64_transpose_interleave_16_1x8.hpp" +#include "a64_transpose_interleave_16_2x2.hpp" +#include "a64_transpose_interleave_16_2x4.hpp" +#include "a64_transpose_interleave_16_2x4_fp32bf16.hpp" +#include "a64_transpose_interleave_16.hpp" +#include "a64_transpose_interleave_24_bf16fp32.hpp" +#include "a64_transpose_interleave_24_fp16fp32.hpp" +#include "a64_transpose_interleave_24_2x4_fp32bf16.hpp" +#include "a64_transpose_interleave_24.hpp" +#include "a64_transpose_interleave_32_1x4.hpp" +#include "a64_transpose_interleave_32_2x2.hpp" +#include "a64_transpose_interleave_4_1x16.hpp" +#include "a64_transpose_interleave_4_1x4.hpp" +#include "a64_transpose_interleave_48.hpp" +#include "a64_transpose_interleave_64.hpp" +#include "a64_transpose_interleave_96.hpp" |