mirror/opencv
1
0
Fork 0
mirror of https://github.com/opencv/opencv.git synced 2024-11-24 03:00:14 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

Use LMUL=2 in the RISC-V Vector (RVV) FP16 part.

Browse Source
This commit is contained in:
Liutong HAN 2024-10-24 15:08:43 +08:00
parent a95658f106
commit a59a66a2c7
2 changed files with 432 additions and 472 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

899
modules/core/include/opencv2/core/hal/intrin_rvv_scalable.hpp
View File

File diff suppressed because it is too large Load Diff

5
modules/core/src/matmul.simd.hpp
View File

@ -1595,7 +1595,10 @@ transform_16u( const ushort* src, ushort* dst, const float* m, int len, int scn,
static void
transform_32f( const float* src, float* dst, const float* m, int len, int scn, int dcn )
{
#if (CV_SIMD || CV_SIMD_SCALABLE) && !defined(__aarch64__) && !defined(_M_ARM64)
// Disabled for RISC-V Vector (scalable), because of:
// 1. v_matmuladd for RVV is 128-bit only but not scalable, this will fail the test `Core_Transform.accuracy`.
// 2. Both gcc and clang can autovectorize this, with better performance than using Universal intrinsic.
#if (CV_SIMD || CV_SIMD_SCALABLE) && !defined(__aarch64__) && !defined(_M_ARM64) && !(CV_TRY_RVV && CV_RVV)
int x = 0;
if( scn == 3 && dcn == 3 )
{