SBR DSP x86: implement SSE sbr_hf_g_filt

Unrolling the main loop to process, instead of 4 elements:
- 8: minor gain of 2 cycles (not worth the extra object size)
- 2: loss of 8 cycles.

Assigning STEP to a register is a loss. Output address (Y) is almost always
unaligned.

Timings:
- C (32/64 bits): 117/109 cycles
- SSE: 57 cycles

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>

1 files changed, 3 insertions, 0 deletions

diff --git a/libavcodec/x86/sbrdsp_init.c b/libavcodec/x86/sbrdsp_init.c
index 313f492054..0ffe5b9e11 100644
--- a/libavcodec/x86/sbrdsp_init.c
+++ b/libavcodec/x86/sbrdsp_init.c
@@ -24,6 +24,8 @@
 #include "libavcodec/sbrdsp.h"
 
 float ff_sbr_sum_square_sse(float (*x)[2], int n);
+void ff_sbr_hf_g_filt_sse(float (*Y)[2], const float (*X_high)[40][2],
+                          const float *g_filt, int m_max, intptr_t ixh);
 
 void ff_sbrdsp_init_x86(SBRDSPContext *s)
 {
@@ -32,6 +34,7 @@ void ff_sbrdsp_init_x86(SBRDSPContext *s)
 
         if (mm_flags & AV_CPU_FLAG_SSE) {
             s->sum_square = ff_sbr_sum_square_sse;
+            s->hf_g_filt = ff_sbr_hf_g_filt_sse;
         }
     }
 }