x86/flacdsp: add a SSE2 version of wasted32

flac_wasted_32_c: 851.3 flac_wasted_32_sse2: 41.3

James Almer committed 2 years ago

Commit 0664cbd1 parent 5acec182 files · +27 −0Browse files

Showing the changes to libavcodec (2 of 2 files). Show all

Unified · Split

2 files

@@ -89,6 +89,30 @@ LPC_32 sse4, 32, psrlq · 88 unmodified lines
89	LPC_32 xop, 32, psrlq	89	LPC_32 xop, 32, psrlq
90	%endif	90	%endif
91		91
		92	INIT_XMM sse2
		93	cglobal flac_wasted_32, 3,3,5, decoded, wasted, len
		94	shl lend, 2
		95	add decodedq, lenq
		96	neg lenq
		97	movd m4, wastedd
		98	ALIGN 16
		99	.loop:
		100	mova m0, [decodedq+lenq+mmsize*0]
		101	mova m1, [decodedq+lenq+mmsize*1]
		102	mova m2, [decodedq+lenq+mmsize*2]
		103	mova m3, [decodedq+lenq+mmsize*3]
		104	pslld m0, m4
		105	pslld m1, m4
		106	pslld m2, m4
		107	pslld m3, m4
		108	mova [decodedq+lenq+mmsize*0], m0
		109	mova [decodedq+lenq+mmsize*1], m1
		110	mova [decodedq+lenq+mmsize*2], m2
		111	mova [decodedq+lenq+mmsize*3], m3
		112	add lenq, mmsize * 4
		113	jl .loop
		114	RET
		115
92	;----------------------------------------------------------------------------------	116	;----------------------------------------------------------------------------------
93	;void ff_flac_decorrelate_[lrm]s_16_sse2(uint8_t out, int32_t in, int channels,	117	;void ff_flac_decorrelate_[lrm]s_16_sse2(uint8_t out, int32_t in, int channels,
94	; int len, int shift);	118	; int len, int shift);

 void ff_flac_lpc_32_xop(int32_t *samples, const int coeffs[32], int order,
                         int qlevel, int len);
+void ff_flac_wasted_32_sse2(int32_t *decoded, int wasted, int len);
 #define DECORRELATE_FUNCS(fmt, opt)                                                      \
 void ff_flac_decorrelate_ls_##fmt##_##opt(uint8_t **out, int32_t **in, int channels,     \
                                           int len, int shift);                           \
     int cpu_flags = av_get_cpu_flags();
     if (EXTERNAL_SSE2(cpu_flags)) {
+        c->wasted32 = ff_flac_wasted_32_sse2;
         if (fmt == AV_SAMPLE_FMT_S16) {
             c->decorrelate[1] = ff_flac_decorrelate_ls_16_sse2;
             c->decorrelate[2] = ff_flac_decorrelate_rs_16_sse2;

Keyboard shortcuts