Mercurial > libavcodec.hg
graph
-
write cabac low and range variables as early as possible to prevent stalls from reading them before they where written, the P4 is said to disslike that alot, on P3 its 2% faster (START/STOP_TIMER over decode_residual) libavcodecWed, 11 Oct 2006 16:11:41 +0000, by michael
-
use ecx instead of cl (no speed change on P3 but might avoid partial register stalls on some cpus) libavcodecWed, 11 Oct 2006 15:20:08 +0000, by michael
-
make state transition tables global as they are constant and the code is slightly faster that way libavcodecWed, 11 Oct 2006 14:44:17 +0000, by michael
-
make lps_range a global table its constant anyway (saves 1 addition for accessing it) libavcodecWed, 11 Oct 2006 13:21:42 +0000, by michael
-
enable CMOV_IS_FAST as its faster or equal speed on every cpu (duron, athlon, PM, P3) from which ive seen benchmarks, it might be slower on P4 but noone has posted benchmarks ... libavcodecWed, 11 Oct 2006 12:23:40 +0000, by michael
-
Switch to the LGPL as agreed to by the author according to the libavcodecWed, 11 Oct 2006 07:47:59 +0000, by diego
-
moving another bit&1 out, this is as fast as with it in there, but it makes more sense with it outside of the loop libavcodecTue, 10 Oct 2006 06:56:51 +0000, by michael
-
move the &1 out of the asm so gcc can optimize it away in inlined cases (yes this is slightly faster) libavcodecTue, 10 Oct 2006 01:17:39 +0000, by michael
-
reading 8bit mem into a 8bit register needs 2 uops on P4, 8bit->32bit with zero extension needs just 1 libavcodecMon, 09 Oct 2006 21:57:10 +0000, by michael
-
reverse remainder of the failed attempt to optimize *state=c->mps_state[s] libavcodecMon, 09 Oct 2006 21:14:16 +0000, by michael
-
removing ALTIVEC_USE_REFERENCE_C_CODE, since has no use anymore libavcodecMon, 09 Oct 2006 18:29:46 +0000, by lu_zero
-
move outcommented START/STOP_TIMER to a hopefully better place for benchmarking ... libavcodecMon, 09 Oct 2006 18:20:00 +0000, by michael
-
drop failed attempt to optimize *state= c->mps_state[s]; libavcodecMon, 09 Oct 2006 15:52:17 +0000, by michael
-
first try of a handwritten get_cabac() for x86, this is 10-20% faster on P3 depening on if you try to subtract the START/STOP_TIMER overhead libavcodecMon, 09 Oct 2006 14:15:14 +0000, by michael
-
remove bytestream_end checks, seems to work fine without them and the bitstream reader doesnt check for the end either libavcodecMon, 09 Oct 2006 12:25:24 +0000, by michael