Date: Sun, 11 Apr 2004 13:09:44 +0200
From: Stephane Marchesin
Subject: [SDL] [Patch] inlining memcpy functions
I (finally) did some benchmarking of the misc mmx & sse blitting
functions, and found a little bottleneck in the memcpy ones : you get
~10% more performance on small surface blitting if you inline them.
--- a/src/video/SDL_blit.c Sun Apr 11 19:49:34 2004 +0000
+++ b/src/video/SDL_blit.c Sun Apr 11 19:52:58 2004 +0000
@@ -116,7 +116,7 @@
}
#ifdef MMX_ASMBLIT
-void SDL_memcpyMMX(char* to,char* from,int len)
+static __inline__ void SDL_memcpyMMX(char* to,char* from,int len)
{
int i;
@@ -132,7 +132,7 @@
SDL_memcpy(to, from, len&7);
}
-void SDL_memcpySSE(char* to,char* from,int len)
+static __inline__ void SDL_memcpySSE(char* to,char* from,int len)
{
int i;