Commit Graph

94 Commits

Author SHA1 Message Date
Yunqing Wang
fe270dd527 Specify size for argument pushed to stack
The change fixes building error on Win64.

Change-Id: I63d25b26220c4da8a98ca2e36530cbb802468e6b
2011-07-25 11:30:45 -04:00
Yunqing Wang
20bd1446c0 Preload reference area to an intermediate buffer in sub-pixel motion search
In sub-pixel motion search, the search range is small(+/- 3 pixels).
Preload whole search area from reference buffer into a 32-byte
aligned buffer. Then in search, load reference data from this buffer
instead. This keeps data in cache, and reduces the crossing cache-
line penalty. For tulip clip, tests on Intel Core2 Quad machine(linux)
showed encoder speed improvement:
  3.4%   at --rt --cpu-used =-4
  2.8%   at --rt --cpu-used =-3
  2.3%   at --rt --cpu-used =-2
  2.2%   at --rt --cpu-used =-1

Test on Atom notebook showed only 1.1% speed improvement(speed=-4).
Test on Xeon machine also showed less improvement, since unaligned
data access latency is greatly reduced in newer cores.

Next, I will apply similar idea to other 2 sub-pixel search functions
for encoding speed > 4.

Make this change exclusively for x86 platforms.

Change-Id: Ia7bb9f56169eac0f01009fe2b2f2ab5b61d2eb2f
2011-07-22 09:28:06 -04:00
Johann
92b0e544f3 fix --disable-runtime-cpu-detect on x86
Change-Id: Ib8e429152c9a8b6032be22b5faac802aa8224caa
2011-06-14 11:31:50 -04:00
Yaowu Xu
361717d2be remove one set of 16x16 variance funcations
call to this set of functions are replaced by var16x16.

Change-Id: I5ff1effc6c1358ea06cda1517b88ec28ef551b0d
2011-06-09 11:23:05 -07:00
Yaowu Xu
d4700731ca remove redundant functions
The encoder defined about 4 set of similar functions to calculate sum,
variance or sse or a combination of them. This commit removed one set
of these functions, get8x8var and get16x16var, where calls to the later
function are replaced with var16x16 by using the fact on a 16x16 MB:
    variance == sse - sum*sum/256

Change-Id: I803eabd1fb3ab177780a40338cbd596dffaed267
2011-06-06 16:44:05 -07:00
Yunqing Wang
b6679879b8 Return sse value in vp8_variance SSE2 functions
Minor modification.

Change-Id: I09511d38fd1451d5c4106a48acdb3f766ce59cb7
2011-05-25 11:55:41 -04:00
John Koleszar
c684d5e5f2 Merge "changed configure option name to reduce confusion" 2011-05-19 11:17:08 -07:00
Yunqing Wang
c7a56f677d Merge "Use diamond search to replace full search in full-pixel refining search" 2011-05-10 06:59:38 -07:00
Yunqing Wang
cb7b1fb144 Use diamond search to replace full search in full-pixel refining search
In NEWMV mode, currently, full search is used as the refining search
after n-step search. By replacing it with an iterative diamond search
of radius 1 largely reduced the computation complexity, but still
maintained the same encoding quality since the refining search is
done for every macroblock instead of only a small precentage of
macroblocks while using full search.

Tests on the test set showed a 3.4% encoding speed increase with none
psnr & ssim loss.

Change-Id: Ife907d7eb9544d15c34f17dc6e4cfd97cb743d41
2011-05-09 14:07:06 -04:00
Johann
a7d4d3c550 clean up unused variable warnings
Change-Id: I9467d7a50eac32d8e8f3a2f26db818e47c93c94b
2011-05-09 12:56:20 -04:00
Yaowu Xu
57ad189129 changed configure option name to reduce confusion
Renamed configure option "enable-psnr" to "enable-internal-stats" to
better reflect the purpose of the option and eliminate the confusion
reported in http://code.google.com/p/webm/issues/detail?id=35

Change-Id: If72df6fdb9f1e33dab1329240ba4d8911d2f1f7a
2011-04-29 09:39:05 -07:00
Johann
aeca599087 Merge "keep values in registers during quantization" 2011-04-25 06:52:38 -07:00
Ronald S. Bultje
496bcbb0de Fix overflow in temporal_filter_apply_sse2().
The accumulator array is an integer array, so use paddd instead of paddw
to add values to it. Fixes overflows when using large --arnr-maxframes
(>8) values.

Change-Id: Iad83794caa02400a65f3ab5760f2517e082d66ae
2011-04-22 10:00:38 -04:00
Johann
508ae1b3d5 keep values in registers during quantization
add an sse4 quantizer so we can use pinsrw/pextrw and keep values in xmm
registers instead of proxying through the stack. and as long as we're
bumping up, use some ssse3 instructions in the EOB detection (see ssse3
fast quantizer)
pick up about a percent on 32bit and about two on 64bit.

Change-Id: If15abba0e8b037a1d231c0edf33501545c9d9363
2011-04-21 15:47:55 -04:00
Johann
4a2b684ef4 modify SAVE_XMM for potential 64bit use
the win64 abi requires saving and restoring xmm6:xmm15. currently
SAVE_XMM and RESTORE XMM only allow for saving xmm6:xmm7. allow
specifying the highest register used and if the stack is unaligned.

Change-Id: Ica5699622ffe3346d3a486f48eef0206c51cf867
2011-04-19 10:42:45 -04:00
Johann
a9b465c5c9 Merge "Add save/restore xmm registers in x86 assembly code" 2011-04-19 06:32:10 -07:00
Johann
c7cfde42a9 Add save/restore xmm registers in x86 assembly code
Went through the code and fixed it. Verified on Windows.

Where possible, remove dependencies on xmm[67]

Current code relies on pushing rbp to the stack to get 16 byte
alignment. This broke when rbp wasn't pushed
(vp8/encoder/x86/sad_sse3.asm). Work around this by using unaligned
memory accesses. Revisit this and the offsets in
vp8/encoder/x86/sad_sse3.asm in another change to SAVE_XMM.

Change-Id: I5f940994d3ebfd977c3d68446cef20fd78b07877
2011-04-18 16:30:38 -04:00
Johann
cd103a5721 Merge "store quant_shift as an unsigned char" 2011-04-18 10:03:40 -07:00
Yaowu Xu
c619f6cb0f Merge "fixed an overflow in ssim calculation" 2011-04-18 07:44:34 -07:00
Johann
70f30aa95d store quant_shift as an unsigned char
in encodframe.c, quant_shift is set to 0 or 1 in vp8cx_invert_quant

only use 8 bits to store this, instead of 16. will allow saving an
xmm register in an updated version of the regular quantize

Change-Id: Ie88c47fe2aff5af0283dab1147fb2791e4b12f90
2011-04-13 13:50:12 -04:00
Yunqing Wang
4fd81a99f8 Set cpu_used range to [-16, 16] in real-time mode
Remove encoding speed limitation in real-time mode.

Change-Id: Ib5e35d8bb522b2a25f3e4ad5cfe2788ebebb3617
2011-04-11 15:55:04 -04:00
Jim Bankoski
d4cdb683a4 fixed an overflow in ssim calculation
This commit fixed an overflow in ssim calculation, added register
save and restore to make sure assembly code working for x64 platform.
It also changed the sampling points to every 4x4 instead of 8x8 and
adjusted the constants in SSIM calculation to match the scale of
previous VPXSSIM.

Change-Id: Ia4dbb8c69eac55812f4662c88ab4653b6720537b
2011-04-07 14:25:25 -07:00
Johann Koenig
08702002e8 use asm_offsets with vp8_fast_quantize_b_sse3
on the same order as the sse2 fast quantize change: ~2%
except for 32bit. only a slight improvment there.

Change-Id: Iff80e5f1ce7e646eebfdc8871405458ff911986b
2011-04-07 16:40:05 -04:00
James Berry
aec5487cdd Use correct 32 bit comparisons for SAD breakout.
Rax updated to eax to avoid uninitialized memory
usage.

Change-Id: Iedb953f104329ede2a786fc648a47f1be2f3798a
2011-04-07 15:08:03 -04:00
Johann
c32e0ecc59 use asm_offsets with vp8_fast_quantize_b_sse2
on the same order as the regular quantize change: ~2%

Change-Id: I5c9eec18e89ae7345dd96945cb740e6f349cee86
2011-04-04 16:23:29 -04:00
Johann
8520b5c785 tweak vp8_regular_quantize_b_sse2
rather than look up rc in the zig zag table, embed it in the macro. this
also allows us to shuffle some values in the macro and keep *d in rsi

gains of about the same order as the obj_int_extract implementation: ~2%

Change-Id: Ib7252dd10eee66e0af8b0e567426122781dc053d
2011-04-01 09:58:23 -04:00
Yunqing Wang
534ea700bd Merge "Fix a crash while enabling shared (--enable-shared)" 2011-03-29 09:04:22 -07:00
Yunqing Wang
b843aa4eda Fix a crash while enabling shared (--enable-shared)
Fixed a bug in SSSE3 sub-pixel filter functions.

Change-Id: I2e2126652970eb78307ffcefcace1efd5966fb0a
2011-03-29 11:31:06 -04:00
Johann
f0c22a3f33 use GLOBAL correctly on 32bit shared libraries
http://code.google.com/p/webm/issues/detail?id=309

Change-Id: I6fce9e2f74bc09a9f258df7f91ab599812324e8c
2011-03-29 11:27:03 -04:00
Johann
8edaf6e2f2 use asm_offsets with vp8_regular_quantize_b_sse2
remove helper function and avoid shadowing all the arguments to the
stack on 64bit systems

when running with --good --cpu-used=0:
~2% on linux x86 and x86_64
~2% on win32 x86 msys and visual studio
more on darwin10 x86_64
significantly more on
x86_64-win64-vs9

Change-Id: Ib7be12edf511fbf2922f191afd5b33b19a0c4ae6
2011-03-24 13:34:48 -04:00
John Koleszar
2cbd962088 Remove unused vp8_get4x4sse_cs_mmx declaration
This declaration did not match the prototype_sad() prototype, but was
unused in this translation unit, so it is removed instead. Fixes
issue 290.

Change-Id: I168854f88a85f73ca9aaf61d1e5dc0f43fc3fdb3
2011-03-21 07:53:53 -04:00
John Koleszar
429dc676b1 Increase static linkage, remove unused functions
A large number of functions were defined with external linkage, even
though they were only used from within one file. This patch changes
their linkage to static and removes the vp8_ prefix from their names,
which should make it more obvious to the reader that the function is
contained within the current translation unit. Functions that were
not referenced were removed.

These symbols were identified by:

  $ nm -A libvpx.a | sort -k3 | uniq -c -f2 | grep ' [A-Z] ' \
    | sort | grep '^ *1 '

Change-Id: I59609f58ab65312012c047036ae1e0634f795779
2011-03-17 20:53:47 -04:00
Yunqing Wang
3c9dd6c3ef Merge "Align SAD output array to be 16-byte aligned" 2011-03-11 05:56:02 -08:00
Jim Bankoski
3f6f7289aa vp8cx- alternate ssim function with optimizations
Change-Id: I91921b0a90dbaddc7010380b038955be347964b3
2011-03-11 08:51:21 -05:00
Yunqing Wang
b2aa401776 Align SAD output array to be 16-byte aligned
Use aligned store.

Change-Id: Icab4c0c53da811d0c52bb7e8134927f249ba2499
2011-03-11 08:24:23 -05:00
Yunqing Wang
7b8e7f0f3a Add vp8_sub_pixel_variance16x8_ssse3 function
Added SSSE3 function

Change-Id: I8c304c92458618d93fda3a2f62bd09ccb63e75ad
2011-03-09 12:33:21 -05:00
Yunqing Wang
4561109a69 Remove unused functions
Removed some unused functions

Change-Id: Ifdfc27453e53cfc75997b38492901d193a16b245
2011-03-09 10:45:03 -05:00
Yunqing Wang
419f638910 Improve SSE2 half-pixel filter funtions
Rewrote these functions to process 16 pixels once instead of 8.

Change-Id: Ic67e80124467a446a3df4cfecfb76a4248602adb
2011-03-08 16:25:06 -05:00
Yunqing Wang
8432a1729f Add zero offset checking in SSE2 sub-pixel filter function
Skip filter at zero offset.

Change-Id: I95fc7e211869bc0ab5bcfb7ab2e3259d1c0ccf38
2011-03-08 15:22:07 -05:00
Yunqing Wang
244e2e1451 Write SSSE3 sub-pixel filter function
1. Process 16 pixels at one time instead of 8.
2. Add check for both xoffset =0 and yoffset=0, which happens
   during motion search.
This change gave encoder 1%~3% performance gain.

Change-Id: Idaa39506b48f4f8b2fbbeb45aae8226fa32afb3e
2011-03-08 13:29:01 -05:00
Yunqing Wang
cfaee9f7c6 Merge "Add prefetch before variance calculation" 2011-02-28 11:42:28 -08:00
Yunqing Wang
d96ba65a23 Add prefetch before variance calculation
This improved encoding performance by 0.5% (good, speed 1) to
1.5% (good, speed 5).

Change-Id: I843d72a0d68a90b5f694adf770943e4a4618f50e
2011-02-28 11:25:55 -05:00
Attila Nagy
7af0d906e3 Remove temporal alt ref from realtime only build
It is not used in realtime mode. Reduces memory footprint.

Change-Id: I7f163225762368df5457cfd413050161d3704a3f
2011-02-22 12:53:32 +02:00
Johann
945dad277d Revert "use unaligned load"
This reverts commit f50f2fd2a7.

Change Ib7506e3e aligns the buffer

Change-Id: Ie0f8bd3e57cfdfef81d39638a1451458ebbae2e0
2011-02-18 10:23:02 -05:00
John Koleszar
c351aa7f1b Merge "Fix relative include paths" 2011-02-17 04:13:44 -08:00
Yunqing Wang
2debd5b5f7 Improve vp8_sad16x16_sse3 function
In real-time mode, vp8_sad16x16 function is called heavily in
motion search part. Improvement of this function gives 1.2%
encoding performance gain (real-time mode, tulip clip).

Change-Id: I23c401fc40c061f732a9767e8d383737a179bd58
2011-02-14 16:23:49 -05:00
John Koleszar
02321de0f2 Fix relative include paths
Allow compiling without adding vp8/{common,encoder,decoder} to the
include paths.

Change-Id: Ifeb5dac351cdfadcd659736f5158b315a0030b6c
2011-02-10 15:09:44 -05:00
Johann
907e98fbb5 Merge "update sse2 regular quantizer" 2011-01-25 13:40:28 -08:00
Yunqing Wang
0822a62f40 Modify sub-pixel filters to eliminate unnecessary calculations
In sub-pixel calculation, xoffset and yoffset mostly take some
specific values. Modified sub-pixel filter functions according to
these possible values to improve performance.

Change-Id: I83083570af8b00ff65093467914fbb97a4e9ea21
2011-01-21 13:59:27 -05:00
Attila Nagy
cb791aaa2f Fix encoder real-time only configuration.
Remove allocation/deallocation of stats storage.
Remove full search functions in machine specific encoder inits.
Remove last pass validation in  validate_config.

Change-Id: I7f29be69273981a4fef6e80ecdb6217c68cbad4e
2011-01-18 08:19:21 -05:00