Mercurial > libavcodec.hg
comparison x86/Makefile @ 12492:58a960d6e34c libavcodec
Rename h264_idct_sse2.asm to h264_idct.asm; move inline IDCT asm from
h264dsp_mmx.c to h264_idct.asm (as yasm code). Because the loops are now
coded in asm instead of C, this is (depending on the function) up to 50%
faster for cases where gcc didn't do a great job at looping.
Since h264_idct_add8() is now faster than the manual loop setup in h264.c,
in-asm idct calling can now be enabled for chroma as well (see r16207). For
MMX, this is 5% faster. For SSE2 (which isn't done for chroma if h264.c does
the looping), this makes it up to 50% faster. Speed gain overall is ~0.5-1.0%.
| author | rbultje |
|---|---|
| date | Tue, 14 Sep 2010 13:36:26 +0000 |
| parents | 14896fa76003 |
| children | c5ffa8b81f9c |
comparison
equal
deleted
inserted
replaced
| 12491:990f8a5fc8af | 12492:58a960d6e34c |
|---|---|
| 8 $(YASM-OBJS-FFT-yes) | 8 $(YASM-OBJS-FFT-yes) |
| 9 | 9 |
| 10 MMX-OBJS-$(CONFIG_H264DSP) += x86/h264dsp_mmx.o | 10 MMX-OBJS-$(CONFIG_H264DSP) += x86/h264dsp_mmx.o |
| 11 YASM-OBJS-$(CONFIG_H264DSP) += x86/h264_deblock.o \ | 11 YASM-OBJS-$(CONFIG_H264DSP) += x86/h264_deblock.o \ |
| 12 x86/h264_weight.o \ | 12 x86/h264_weight.o \ |
| 13 x86/h264_idct_sse2.o \ | 13 x86/h264_idct.o \ |
| 14 | 14 |
| 15 YASM-OBJS-$(CONFIG_H264PRED) += x86/h264_intrapred.o | 15 YASM-OBJS-$(CONFIG_H264PRED) += x86/h264_intrapred.o |
| 16 MMX-OBJS-$(CONFIG_H264PRED) += x86/h264_intrapred_init.o | 16 MMX-OBJS-$(CONFIG_H264PRED) += x86/h264_intrapred_init.o |
| 17 | 17 |
| 18 YASM-OBJS-$(CONFIG_VC1_DECODER) += x86/vc1dsp_yasm.o | 18 YASM-OBJS-$(CONFIG_VC1_DECODER) += x86/vc1dsp_yasm.o |