X-Git-Url: https://bilbo.iut-bm.univ-fcomte.fr/and/gitweb/book_gpu.git/blobdiff_plain/1b4c95e988aa2d0c5d7f6cf87750ea4214dc8171..dd9c1efaf5eab342643f363e8bd4ab8012ba9596:/BookGPU/Chapters/chapter4/code/convoGeneSh1.cu?ds=inline

diff --git a/BookGPU/Chapters/chapter4/code/convoGeneSh1.cu b/BookGPU/Chapters/chapter4/code/convoGeneSh1.cu
index 53d9ba4..95eed34 100644
--- a/BookGPU/Chapters/chapter4/code/convoGeneSh1.cu
+++ b/BookGPU/Chapters/chapter4/code/convoGeneSh1.cu
@@ -11,7 +11,7 @@ __global__ void kernel_convoNonSepSh_8p(unsigned char *output, int j_dim, int r)
   int i = __umul24( blockIdx.y, blockDim.y) + threadIdx.y ;
   int j0= __umul24(blockIdx.x,blockDim.x)<<3 ;   // block's base point
   int idx = __umul24(i,j_dim) + j ;              // absolute index
-  int idrow = threadIdx.y*(bdimX+k-1) ;          // line's offset in sh mem
+  int idrow = threadIdx.y*(bdimX+k-1) ;    // line's offset in sh mem
   
   // shared memory declaration
   extern __shared__ unsigned char roi8p[];
@@ -46,7 +46,7 @@ __global__ void kernel_convoNonSepSh_8p(unsigned char *output, int j_dim, int r)
 	for( jc=0 ; jc<k ; jc++)
 	  {
 		int baseRoi = __umul24(ic+threadIdx.y,(bdimX+k-1)) + jc+tidX ;
-		float valMask = masque[ __umul24(ic,k)+jc ] ;
+		float valMask = mask[ __umul24(ic,k)+jc ] ;
 		outval0 += valMask*roi8p[ baseRoi ] ;
 		outval1 += valMask*roi8p[ baseRoi +1 ] ;
 		outval2 += valMask*roi8p[ baseRoi +2 ] ;
@@ -65,5 +65,5 @@ __global__ void kernel_convoNonSepSh_8p(unsigned char *output, int j_dim, int r)
   output[ idx++ ] = outval4 ;
   output[ idx++ ] = outval5 ;
   output[ idx++ ] = outval6 ;
-  output[ idx++ ] = outval7 ;
+  output[ idx   ] = outval7 ;
 }