potok/opencv
1
0
Fork 0
mirror of https://github.com/opencv/opencv.git synced 2026-07-31 08:13:04 +04:00
Code Packages Releases Wiki Activity

Merge pull request #27050 from hanliutong:rvv-fix-27003

Browse Source 
RISC-V: Fix #27003.
This commit is contained in:
Alexander Smorkalov
2025-03-12 17:32:19 +03:00
committed by GitHub
parent eefa327f30 2969b67bd7
commit 49ab8121b7
1 changed files with 11 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
modules/core/include/opencv2/core/hal/intrin_rvv_scalable.hpp
+11 -1
View File
@@ -2130,7 +2130,17 @@ inline v_int64 v_dotprod_expand_fast(const v_int16& a, const v_int16& b, const v
// 32 >> 64f
#if CV_SIMD_SCALABLE_64F
inline v_float64 v_dotprod_expand_fast(const v_int32& a, const v_int32& b)
{ return v_cvt_f64(v_dotprod_fast(a, b)); }
{
vfloat64m1_t zero = __riscv_vfmv_v_f_f64m1(0, VTraits<vuint64m1_t>::vlanes());
auto prod_i64 = __riscv_vwmul(a, b, VTraits<v_int32>::vlanes());
// Convert to f64 before reduction to avoid overflow: #27003
auto prod_f64 = __riscv_vfcvt_f(prod_i64, VTraits<v_int32>::vlanes());
return __riscv_vset( // Needs v_float64 (vfloat64m2_t) here.
v_setall_f64(0.0f), // zero_f64m2
0,
__riscv_vfredusum_tu(zero, prod_f64, zero, VTraits<v_int32>::vlanes())
);
}
inline v_float64 v_dotprod_expand_fast(const v_int32& a, const v_int32& b, const v_float64& c)
{ return v_add(v_dotprod_expand_fast(a, b) , c); }
#endif