Revision 8a322796 libswscale/yuv2rgb_altivec.c

View differences:

libswscale/yuv2rgb_altivec.c
21 21
 */
22 22

  
23 23
/*
24
convert I420 YV12 to RGB in various formats,
25
  it rejects images that are not in 420 formats
26
  it rejects images that don't have widths of multiples of 16
27
  it rejects images that don't have heights of multiples of 2
28
reject defers to C simulation codes.
24
Convert I420 YV12 to RGB in various formats,
25
  it rejects images that are not in 420 formats,
26
  it rejects images that don't have widths of multiples of 16,
27
  it rejects images that don't have heights of multiples of 2.
28
Reject defers to C simulation code.
29 29

  
30
lots of optimizations to be done here
30
Lots of optimizations to be done here.
31 31

  
32
1. need to fix saturation code, I just couldn't get it to fly with packs and adds.
33
   so we currently use max min to clip
32
1. Need to fix saturation code. I just couldn't get it to fly with packs
33
   and adds, so we currently use max/min to clip.
34 34

  
35
2. the inefficient use of chroma loading needs a bit of brushing up
35
2. The inefficient use of chroma loading needs a bit of brushing up.
36 36

  
37
3. analysis of pipeline stalls needs to be done, use shark to identify pipeline stalls
37
3. Analysis of pipeline stalls needs to be done. Use shark to identify
38
   pipeline stalls.
38 39

  
39 40

  
40 41
MODIFIED to calculate coeffs from currently selected color space.
41
MODIFIED core to be a macro which you spec the output format.
42
ADDED UYVY conversion which is never called due to some thing in SWSCALE.
42
MODIFIED core to be a macro where you specify the output format.
43
ADDED UYVY conversion which is never called due to some thing in swscale.
43 44
CORRECTED algorithim selection to be strict on input formats.
44
ADDED runtime detection of altivec.
45
ADDED runtime detection of AltiVec.
45 46

  
46 47
ADDED altivec_yuv2packedX vertical scl + RGB converter
47 48

  
48 49
March 27,2004
49 50
PERFORMANCE ANALYSIS
50 51

  
51
The C version use 25% of the processor or ~250Mips for D1 video rawvideo used as test
52
The ALTIVEC version uses 10% of the processor or ~100Mips for D1 video same sequence
52
The C version uses 25% of the processor or ~250Mips for D1 video rawvideo
53
used as test.
54
The AltiVec version uses 10% of the processor or ~100Mips for D1 video
55
same sequence.
53 56

  
54
720*480*30  ~10MPS
57
720 * 480 * 30  ~10MPS
55 58

  
56
so we have roughly 10clocks per pixel this is too high something has to be wrong.
59
so we have roughly 10 clocks per pixel. This is too high, something has
60
to be wrong.
57 61

  
58
OPTIMIZED clip codes to utilize vec_max and vec_packs removing the need for vec_min.
62
OPTIMIZED clip codes to utilize vec_max and vec_packs removing the
63
need for vec_min.
59 64

  
60
OPTIMIZED DST OUTPUT cache/dma controls. we are pretty much
61
guaranteed to have the input video frame it was just decompressed so
62
it probably resides in L1 caches.  However we are creating the
63
output video stream this needs to use the DSTST instruction to
64
optimize for the cache.  We couple this with the fact that we are
65
not going to be visiting the input buffer again so we mark it Least
66
Recently Used.  This shaves 25% of the processor cycles off.
65
OPTIMIZED DST OUTPUT cache/DMA controls. We are pretty much guaranteed to have
66
the input video frame, it was just decompressed so it probably resides in L1
67
caches. However, we are creating the output video stream. This needs to use the
68
DSTST instruction to optimize for the cache. We couple this with the fact that
69
we are not going to be visiting the input buffer again so we mark it Least
70
Recently Used. This shaves 25% of the processor cycles off.
67 71

  
68
Now MEMCPY is the largest mips consumer in the system, probably due
72
Now memcpy is the largest mips consumer in the system, probably due
69 73
to the inefficient X11 stuff.
70 74

  
71 75
GL libraries seem to be very slow on this machine 1.33Ghz PB running
72 76
Jaguar, this is not the case for my 1Ghz PB.  I thought it might be
73
a versioning issues, however I have libGL.1.2.dylib for both
74
machines. ((We need to figure this out now))
77
a versioning issue, however I have libGL.1.2.dylib for both
78
machines. (We need to figure this out now.)
75 79

  
76
GL2 libraries work now with patch for RGB32
80
GL2 libraries work now with patch for RGB32.
77 81

  
78
NOTE quartz vo driver ARGB32_to_RGB24 consumes 30% of the processor
82
NOTE: quartz vo driver ARGB32_to_RGB24 consumes 30% of the processor.
79 83

  
80
Integrated luma prescaling adjustment for saturation/contrast/brightness adjustment.
84
Integrated luma prescaling adjustment for saturation/contrast/brightness
85
adjustment.
81 86
*/
82 87

  
83 88
#include <stdio.h>

Also available in: Unified diff