ardour/livetrax
13
0
Fork 0
Code

Add Mike's quad-quad + prefetch optimization.

Browse Source 
git-svn-id: svn://localhost/ardour2/trunk@1597 d708f5d6-7413-0410-9779-e7cbd77b26cf
This commit is contained in:
Sampo Savolainen 2007-03-15 19:54:18 +00:00
parent 76afeee82b
commit 52c8e00edd
1 changed files with 23 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

23
libs/ardour/sse_functions_xmm.cc
View File

@ -43,6 +43,29 @@ x86_sse_find_peaks(float *buf, nframes_t nframes, float *min, float *max)
nframes--;
}
// use 64 byte prefetch for quadruple quads
while (nframes >= 16) {
__builtin_prefetch(buf+64,0,0);
work = _mm_load_ps(buf);
current_min = _mm_min_ps(current_min, work);
current_max = _mm_max_ps(current_max, work);
buf+=4;
work = _mm_load_ps(buf);
current_min = _mm_min_ps(current_min, work);
current_max = _mm_max_ps(current_max, work);
buf+=4;
work = _mm_load_ps(buf);
current_min = _mm_min_ps(current_min, work);
current_max = _mm_max_ps(current_max, work);
buf+=4;
work = _mm_load_ps(buf);
current_min = _mm_min_ps(current_min, work);
current_max = _mm_max_ps(current_max, work);
buf+=4;
nframes-=16;
}
// work through aligned buffers
while (nframes >= 4) {