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>

libavcodec/x86/sbrdsp.asm

@@ -72,3 +72,43 @@ cglobal sbr_sum_square, 2, 3, 6
     fld         dword r0m
 %endif
     RET
+
+%define STEP  40*4*2
+cglobal sbr_hf_g_filt, 5, 6, 5
+    lea         r1, [r1 + 8*r4] ; offset by ixh elements into X_high
+    mov         r5, r3
+    and         r3, 0xFC
+    lea         r2, [r2 + r3*4]
+    lea         r0, [r0 + r3*8]
+    neg         r3
+.loop4:
+    movq        m0, [r2 + 4*r3 + 0]
+    movq        m1, [r2 + 4*r3 + 8]
+    movq        m2, [r1 + 0*STEP]
+    movq        m3, [r1 + 2*STEP]
+    movhps      m2, [r1 + 1*STEP]
+    movhps      m3, [r1 + 3*STEP]
+    punpckldq   m0, m0
+    punpckldq   m1, m1
+    mulps       m0, m2
+    mulps       m1, m3
+    movu        [r0 + 8*r3 +  0], m0
+    movu        [r0 + 8*r3 + 16], m1
+    add         r1, 4*STEP
+    add         r3, 4
+    jnz         .loop4
+    and         r5, 3 ; number of single element loops
+    jz          .end
+.loop1: ; element 0 and 1 can be computed at the same time
+    movss       m0, [r2]
+    movq        m2, [r1]
+    punpckldq   m0, m0
+    mulps       m2, m0
+    movq      [r0], m2
+    add         r0, 8
+    add         r2, 4
+    add         r1, STEP
+    dec         r5
+    jnz         .loop1
+.end:
+    RET

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;
         }
     }
 }