Merge pull request #23656 from peters:patch-2

Build fix for AVX 256

Merge pull request #23656 from peters:patch-2
Build fix for AVX 256
4a559bc2 · Alexander Smorkalov · GitHub · 044a3225 · 04970490 · 4a559bc2
隐藏空白更改
内联并排

Showing with 1 addition and 1 deletion

modules/dnn/src/layers/cpu_kernels/convolution.cpp modules/dnn/src/layers/cpu_kernels/convolution.cpp +1 -1

未找到文件。
--- a/modules/dnn/src/layers/cpu_kernels/convolution.cpp
+++ b/modules/dnn/src/layers/cpu_kernels/convolution.cpp
@@ -445,7 +445,7 @@ static inline void packData8(char*& inpbuf, float*& inptrIn, int& in_w, int& x0,
            {
                int k1 = ofstab[k];
 #if CV_SIMD256
-                vx_store(inpbufC_FP32 + k*CONV_NR, vx_load(inptrInC + k1));
+                vx_store(inpbufC_FP32 + k*CONV_NR_FP32, vx_load(inptrInC + k1));
 #elif CV_SIMD128
                v_float32x4 vv0 = v_load(inptrInC + k1);
                v_float32x4 vv1 = v_load(inptrInC + k1 + 4);