Commit 32bac65b authored by Michael Niedermayer's avatar Michael Niedermayer

Merge remote-tracking branch 'qatar/master'

* qatar/master:
  x86: sbrdsp: Implement SSE neg_odd_64

Conflicts:
	libavcodec/x86/sbrdsp.asm
Merged-by: 's avatarMichael Niedermayer <michaelni@gmx.at>
parents 84bfa8be f4b0d12f
...@@ -224,21 +224,21 @@ cglobal sbr_qmf_post_shuffle, 2,3,4,W,z ...@@ -224,21 +224,21 @@ cglobal sbr_qmf_post_shuffle, 2,3,4,W,z
INIT_XMM sse INIT_XMM sse
cglobal sbr_neg_odd_64, 1,2,4,z cglobal sbr_neg_odd_64, 1,2,4,z
lea r1q, [zq+256] lea r1q, [zq+256]
.loop: .loop:
mova m0, [zq+ 0] mova m0, [zq+ 0]
mova m1, [zq+16] mova m1, [zq+16]
mova m2, [zq+32] mova m2, [zq+32]
mova m3, [zq+48] mova m3, [zq+48]
xorps m0, [ps_mask2] xorps m0, [ps_mask2]
xorps m1, [ps_mask2] xorps m1, [ps_mask2]
xorps m2, [ps_mask2] xorps m2, [ps_mask2]
xorps m3, [ps_mask2] xorps m3, [ps_mask2]
mova [zq+ 0], m0 mova [zq+ 0], m0
mova [zq+16], m1 mova [zq+16], m1
mova [zq+32], m2 mova [zq+32], m2
mova [zq+48], m3 mova [zq+48], m3
add zq, 64 add zq, 64
cmp zq, r1q cmp zq, r1q
jne .loop jne .loop
REP_RET REP_RET
...@@ -32,19 +32,19 @@ void ff_sbr_hf_g_filt_sse(float (*Y)[2], const float (*X_high)[40][2], ...@@ -32,19 +32,19 @@ void ff_sbr_hf_g_filt_sse(float (*Y)[2], const float (*X_high)[40][2],
void ff_sbr_hf_gen_sse(float (*X_high)[2], const float (*X_low)[2], void ff_sbr_hf_gen_sse(float (*X_high)[2], const float (*X_low)[2],
const float alpha0[2], const float alpha1[2], const float alpha0[2], const float alpha1[2],
float bw, int start, int end); float bw, int start, int end);
void ff_sbr_qmf_post_shuffle_sse(float W[32][2], const float *z);
void ff_sbr_neg_odd_64_sse(float *z); void ff_sbr_neg_odd_64_sse(float *z);
void ff_sbr_qmf_post_shuffle_sse(float W[32][2], const float *z);
av_cold void ff_sbrdsp_init_x86(SBRDSPContext *s) av_cold void ff_sbrdsp_init_x86(SBRDSPContext *s)
{ {
int mm_flags = av_get_cpu_flags(); int mm_flags = av_get_cpu_flags();
if (EXTERNAL_SSE(mm_flags)) { if (EXTERNAL_SSE(mm_flags)) {
s->neg_odd_64 = ff_sbr_neg_odd_64_sse;
s->sum_square = ff_sbr_sum_square_sse; s->sum_square = ff_sbr_sum_square_sse;
s->sum64x5 = ff_sbr_sum64x5_sse; s->sum64x5 = ff_sbr_sum64x5_sse;
s->hf_g_filt = ff_sbr_hf_g_filt_sse; s->hf_g_filt = ff_sbr_hf_g_filt_sse;
s->hf_gen = ff_sbr_hf_gen_sse; s->hf_gen = ff_sbr_hf_gen_sse;
s->qmf_post_shuffle = ff_sbr_qmf_post_shuffle_sse; s->qmf_post_shuffle = ff_sbr_qmf_post_shuffle_sse;
s->neg_odd_64 = ff_sbr_neg_odd_64_sse;
} }
} }
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment