X-Git-Url: https://bilbo.iut-bm.univ-fcomte.fr/and/gitweb/book_gpu.git/blobdiff_plain/1874c46934f4ba7e8c2013d3829f65309456d292..c08513a856650905e4751c8b52c7cbb661368a5f:/BookGPU/Chapters/chapter4/ch4.tex

diff --git a/BookGPU/Chapters/chapter4/ch4.tex b/BookGPU/Chapters/chapter4/ch4.tex
index 805de25..bd66882 100644
--- a/BookGPU/Chapters/chapter4/ch4.tex
+++ b/BookGPU/Chapters/chapter4/ch4.tex
@@ -240,7 +240,7 @@ However, our technique requires writing one kernel per mask size, which can be s
 \lstinputlisting[label={lst:convoGene8x8pL3},caption=CUDA kernel achieving a $3\times 3$ convolution operation with the mask in symbol memory and direct data fetches in texture memory]{Chapters/chapter4/code/convoGene8x8pL3.cu}
 
 \subsection{Using shared memory to store prefetched data\index{prefetching}.}
- \index{memory~hierarchy!shared~memory}
+ \index{memory hierarchy!shared memory}
 A more convenient way of coding a convolution kernel is to use shared memory to perform a prefetching stage of the whole halo before computing the convolution sums.
 This proves to be quite efficient and more versatile, but it obviously generates some overhead because 
 \begin{itemize}