shithub: openh264

Info  •  Files  •  Log  •  Branches

[Decoder/x86] IDCT one block at a time with SSE2

At lower bitrates, it is overall faster to conditionally do one block
at a time with SSE2 on Haswell and likely other common architectures.
At higher bitrates, it is faster to use the wider routine that IDCTs
four blocks at a time. To avoid potential performance regressions
as compared to MMX, stick with single-block IDCTs with SSE2. There
is still a performance advantage as compared to MMX because the
single-block SSE2 routine is faster than the corresponding MMX
routine.

Stick with four blocks at a time with AVX2 for which that appears
to be consistently faster on Haswell.

--- a/codec/common/x86/dct.asm

+++ b/codec/common/x86/dct.asm

@@ -510,20 +510,11 @@

     ret

 ;***********************************************************************

-; void IdctFourResAddPred_sse2(uint8_t* pPred, int32_t iStride, const int16_t* pDct, const int8_t* pNzc);

-;***********************************************************************

-WELS_EXTERN IdctFourResAddPred_sse2

-    %assign push_num 0

-    LOAD_3_PARA_TO_5_PARA_IDCT

-    jmp prefixed(WelsIDctFourT4Rec_sse2.begin)

-

-;***********************************************************************

 ; void WelsIDctFourT4Rec_sse2(uint8_t *rec, int32_t stride, uint8_t *pred, int32_t pred_stride, int16_t *rs);

 ;***********************************************************************

 WELS_EXTERN WelsIDctFourT4Rec_sse2

     %assign push_num 0

     LOAD_5_PARA

-.begin:

     PUSH_XMM 8

     SIGN_EXTENSION r1, r1d

     SIGN_EXTENSION r3, r3d

--- a/codec/decoder/core/inc/decode_mb_aux.h

+++ b/codec/decoder/core/inc/decode_mb_aux.h

@@ -49,7 +49,6 @@

 void IdctResAddPred_mmx (uint8_t* pPred, const int32_t kiStride, int16_t* pRs);

 void IdctResAddPred_sse2 (uint8_t* pPred, const int32_t kiStride, int16_t* pRs);

 void IdctResAddPred_avx2 (uint8_t* pPred, const int32_t kiStride, int16_t* pRs);

-void IdctFourResAddPred_sse2 (uint8_t* pPred, int32_t iStride, int16_t* pRs, const int8_t* pNzc);

 void IdctFourResAddPred_avx2 (uint8_t* pPred, int32_t iStride, int16_t* pRs, const int8_t* pNzc);

 #endif//X86_ASM

--- a/codec/decoder/core/src/decoder.cpp

+++ b/codec/decoder/core/src/decoder.cpp

@@ -992,7 +992,7 @@

   }

   if (uiCpuFlag & WELS_CPU_SSE2) {

     pCtx->pIdctResAddPredFunc     = IdctResAddPred_sse2;

-    pCtx->pIdctFourResAddPredFunc = IdctFourResAddPred_sse2;

+    pCtx->pIdctFourResAddPredFunc = IdctFourResAddPred_<IdctResAddPred_sse2>;

     pCtx->pGetI16x16LumaPredFunc[I16_PRED_DC] = WelsDecoderI16x16LumaPredDc_sse2;

     pCtx->pGetI16x16LumaPredFunc[I16_PRED_P]  = WelsDecoderI16x16LumaPredPlane_sse2;

--- a/test/decoder/DecUT_IdctResAddPred.cpp

+++ b/test/decoder/DecUT_IdctResAddPred.cpp

@@ -139,7 +139,6 @@

 GENERATE_IDCTRESADDPRED (IdctResAddPred_mmx, WELS_CPU_MMXEXT)

 GENERATE_IDCTRESADDPRED (IdctResAddPred_sse2, WELS_CPU_SSE2)

 GENERATE_IDCTRESADDPRED (IdctResAddPred_avx2, WELS_CPU_AVX2)

-GENERATE_IDCTFOURRESADDPRED (IdctFourResAddPred_sse2, WELS_CPU_SSE2)

 GENERATE_IDCTFOURRESADDPRED (IdctFourResAddPred_avx2, WELS_CPU_AVX2)

 #endif

-- 

⑨