From 5998cf99b08d448dd5158ed6229262aa67ea4a66 Mon Sep 17 00:00:00 2001 From: Wim Taymans Date: Thu, 13 Aug 2009 13:45:01 +0200 Subject: svolume: improve SSE and MMX code --- src/pulsecore/svolume_sse.c | 191 +++++++++++++++++++++++++++++++------------- 1 file changed, 134 insertions(+), 57 deletions(-) (limited to 'src/pulsecore/svolume_sse.c') diff --git a/src/pulsecore/svolume_sse.c b/src/pulsecore/svolume_sse.c index 2d4c541b..ff583a06 100644 --- a/src/pulsecore/svolume_sse.c +++ b/src/pulsecore/svolume_sse.c @@ -99,6 +99,44 @@ pa_volume_ulaw_sse (uint8_t *samples, int32_t *volumes, unsigned channels, unsig } #endif +#define VOLUME_32x16(s,v) /* .. | vh | vl | */ \ + " pxor %%xmm4, %%xmm4 \n\t" /* .. | 0 | 0 | */ \ + " punpcklwd %%xmm4, "#s" \n\t" /* .. | 0 | p0 | */ \ + " pcmpgtw "#s", %%xmm4 \n\t" /* .. | 0 | s(p0) | */ \ + " pand "#v", %%xmm4 \n\t" /* .. | 0 | (vl) | */ \ + " movdqa "#s", %%xmm5 \n\t" \ + " pmulhuw "#v", "#s" \n\t" /* .. | 0 | vl*p0 | */ \ + " psubd %%xmm4, "#s" \n\t" /* .. | 0 | vl*p0 | + sign correct */ \ + " psrld $16, "#v" \n\t" /* .. | p0 | 0 | */ \ + " pmaddwd %%xmm5, "#v" \n\t" /* .. | p0 * vh | */ \ + " paddd "#s", "#v" \n\t" /* .. | p0 * v0 | */ \ + " packssdw "#v", "#v" \n\t" /* .. | p1*v1 | p0*v0 | */ + +#define MOD_ADD(a,b) \ + " add "#a", %3 \n\t" /* channel += inc */ \ + " mov %3, %4 \n\t" \ + " sub "#b", %4 \n\t" /* tmp = channel - channels */ \ + " cmp "#b", %3 \n\t" /* if (channel >= channels) */ \ + " cmovae %4, %3 \n\t" /* channel = tmp */ + +/* swap 16 bits */ +#define SWAP_16(s) \ + " movdqa "#s", %%xmm4 \n\t" /* .. | h l | */ \ + " psrlw $8, %%xmm4 \n\t" /* .. | 0 h | */ \ + " psllw $8, "#s" \n\t" /* .. | l 0 | */ \ + " por %%xmm4, "#s" \n\t" /* .. | l h | */ + +/* swap 2 registers 16 bits for better pairing */ +#define SWAP_16_2(s1,s2) \ + " movdqa "#s1", %%xmm4 \n\t" /* .. | h l | */ \ + " movdqa "#s2", %%xmm5 \n\t" \ + " psrlw $8, %%xmm4 \n\t" /* .. | 0 h | */ \ + " psrlw $8, %%xmm5 \n\t" \ + " psllw $8, "#s1" \n\t" /* .. | l 0 | */ \ + " psllw $8, "#s2" \n\t" \ + " por %%xmm4, "#s1" \n\t" /* .. | l h | */ \ + " por %%xmm5, "#s2" \n\t" + static void pa_volume_s16ne_sse (int16_t *samples, int32_t *volumes, unsigned channels, unsigned length) { @@ -109,25 +147,83 @@ pa_volume_s16ne_sse (int16_t *samples, int32_t *volumes, unsigned channels, unsi * we overread the volume array, which should have enough padding. */ channels = MAX (8, channels); -#define VOLUME_32x16(s,v) /* v1_h | v1_l | v0_h | v0_l */ \ - " pxor %%xmm4, %%xmm4 \n\t" \ - " punpcklwd %%xmm4, "#s" \n\t" /* 0 | p1 | 0 | p0 */ \ - " pcmpgtw "#s", %%xmm4 \n\t" /* select sign from sample */ \ - " pand "#v", %%xmm4 \n\t" /* extract sign correction factors */ \ - " movdqa "#s", %%xmm5 \n\t" \ - " pmulhuw "#v", "#s" \n\t" /* 0 | p1*v1lh | 0 | p0*v0lh */ \ - " psubd %%xmm4, "#s" \n\t" /* sign correction */ \ - " psrld $16, "#v" \n\t" /* 0 | v1h | 0 | v0h */ \ - " pmaddwd %%xmm5, "#v" \n\t" /* p1 * v1h | p0 * v0h */ \ - " paddd "#s", "#v" \n\t" /* p1 * v1 | p0 * v0 */ \ - " packssdw "#v", "#v" \n\t" /* p0*v0 | p1*v1 | p0*v0 | p1*v1 */ + __asm__ __volatile__ ( + " xor %3, %3 \n\t" + " sar $1, %2 \n\t" /* length /= sizeof (int16_t) */ -#define MOD_ADD(a,b) \ - " add "#a", %3 \n\t" \ - " mov %3, %4 \n\t" \ - " sub "#b", %4 \n\t" \ - " cmp "#b", %3 \n\t" \ - " cmovae %4, %3 \n\t" + " test $1, %2 \n\t" /* check for odd samples */ + " je 2f \n\t" + + " movd (%1, %3, 4), %%xmm0 \n\t" /* | v0h | v0l | */ + " movw (%0), %4 \n\t" /* .. | p0 | */ + " movd %4, %%xmm1 \n\t" + VOLUME_32x16 (%%xmm1, %%xmm0) + " movd %%xmm0, %4 \n\t" /* .. | p0*v0 | */ + " movw %4, (%0) \n\t" + " add $2, %0 \n\t" + MOD_ADD ($1, %5) + + "2: \n\t" + " sar $1, %2 \n\t" /* prepare for processing 2 samples at a time */ + " test $1, %2 \n\t" + " je 4f \n\t" + + "3: \n\t" /* do samples in groups of 2 */ + " movq (%1, %3, 4), %%xmm0 \n\t" /* | v1h | v1l | v0h | v0l | */ + " movd (%0), %%xmm1 \n\t" /* .. | p1 | p0 | */ + VOLUME_32x16 (%%xmm1, %%xmm0) + " movd %%xmm0, (%0) \n\t" /* .. | p1*v1 | p0*v0 | */ + " add $4, %0 \n\t" + MOD_ADD ($2, %5) + + "4: \n\t" + " sar $1, %2 \n\t" /* prepare for processing 4 samples at a time */ + " test $1, %2 \n\t" + " je 6f \n\t" + + "5: \n\t" /* do samples in groups of 4 */ + " movdqu (%1, %3, 4), %%xmm0 \n\t" /* | v3h | v3l .. v0h | v0l | */ + " movq (%0), %%xmm1 \n\t" /* .. | p3 .. p0 | */ + VOLUME_32x16 (%%xmm1, %%xmm0) + " movq %%xmm0, (%0) \n\t" /* .. | p3*v3 .. p0*v0 | */ + " add $8, %0 \n\t" + MOD_ADD ($4, %5) + + "6: \n\t" + " sar $1, %2 \n\t" /* prepare for processing 8 samples at a time */ + " cmp $0, %2 \n\t" + " je 8f \n\t" + + "7: \n\t" /* do samples in groups of 8 */ + " movdqu (%1, %3, 4), %%xmm0 \n\t" /* | v3h | v3l .. v0h | v0l | */ + " movdqu 16(%1, %3, 4), %%xmm2 \n\t" /* | v7h | v7l .. v4h | v4l | */ + " movq (%0), %%xmm1 \n\t" /* .. | p3 .. p0 | */ + " movq 8(%0), %%xmm3 \n\t" /* .. | p7 .. p4 | */ + VOLUME_32x16 (%%xmm1, %%xmm0) + VOLUME_32x16 (%%xmm3, %%xmm2) + " movq %%xmm0, (%0) \n\t" /* .. | p3*v3 .. p0*v0 | */ + " movq %%xmm2, 8(%0) \n\t" /* .. | p7*v7 .. p4*v4 | */ + " add $16, %0 \n\t" + MOD_ADD ($8, %5) + " dec %2 \n\t" + " jne 7b \n\t" + "8: \n\t" + + : "+r" (samples), "+r" (volumes), "+r" (length), "=D" (channel), "=&r" (temp) + : "r" ((int64_t)channels) + : "cc" + ); +} + +static void +pa_volume_s16re_sse (int16_t *samples, int32_t *volumes, unsigned channels, unsigned length) +{ + int64_t channel; + int64_t temp; + + /* the max number of samples we process at a time, this is also the max amount + * we overread the volume array, which should have enough padding. */ + channels = MAX (8, channels); __asm__ __volatile__ ( " xor %3, %3 \n\t" @@ -138,9 +234,11 @@ pa_volume_s16ne_sse (int16_t *samples, int32_t *volumes, unsigned channels, unsi " movd (%1, %3, 4), %%xmm0 \n\t" /* do odd sample */ " movw (%0), %4 \n\t" + " rorw $8, %4 \n\t" " movd %4, %%xmm1 \n\t" VOLUME_32x16 (%%xmm1, %%xmm0) " movd %%xmm0, %4 \n\t" + " rorw $8, %4 \n\t" " movw %4, (%0) \n\t" " add $2, %0 \n\t" MOD_ADD ($1, %5) @@ -153,7 +251,9 @@ pa_volume_s16ne_sse (int16_t *samples, int32_t *volumes, unsigned channels, unsi "3: \n\t" /* do samples in pairs of 2 */ " movq (%1, %3, 4), %%xmm0 \n\t" /* v1_h | v1_l | v0_h | v0_l */ " movd (%0), %%xmm1 \n\t" /* X | X | p1 | p0 */ + SWAP_16 (%%xmm1) VOLUME_32x16 (%%xmm1, %%xmm0) + SWAP_16 (%%xmm0) " movd %%xmm0, (%0) \n\t" " add $4, %0 \n\t" MOD_ADD ($2, %5) @@ -164,9 +264,11 @@ pa_volume_s16ne_sse (int16_t *samples, int32_t *volumes, unsigned channels, unsi " je 6f \n\t" "5: \n\t" /* do samples in pairs of 4 */ - " movdqa (%1, %3, 4), %%xmm0 \n\t" /* v1_h | v1_l | v0_h | v0_l */ + " movdqu (%1, %3, 4), %%xmm0 \n\t" /* v1_h | v1_l | v0_h | v0_l */ " movq (%0), %%xmm1 \n\t" /* X | X | p1 | p0 */ + SWAP_16 (%%xmm1) VOLUME_32x16 (%%xmm1, %%xmm0) + SWAP_16 (%%xmm0) " movq %%xmm0, (%0) \n\t" " add $8, %0 \n\t" MOD_ADD ($4, %5) @@ -177,12 +279,14 @@ pa_volume_s16ne_sse (int16_t *samples, int32_t *volumes, unsigned channels, unsi " je 8f \n\t" "7: \n\t" /* do samples in pairs of 8 */ - " movdqa (%1, %3, 4), %%xmm0 \n\t" /* v1_h | v1_l | v0_h | v0_l */ - " movdqa 16(%1, %3, 4), %%xmm2 \n\t" /* v3_h | v3_l | v2_h | v2_l */ + " movdqu (%1, %3, 4), %%xmm0 \n\t" /* v1_h | v1_l | v0_h | v0_l */ + " movdqu 16(%1, %3, 4), %%xmm2 \n\t" /* v3_h | v3_l | v2_h | v2_l */ " movq (%0), %%xmm1 \n\t" /* X | X | p1 | p0 */ " movq 8(%0), %%xmm3 \n\t" /* X | X | p3 | p2 */ + SWAP_16_2 (%%xmm1, %%xmm3) VOLUME_32x16 (%%xmm1, %%xmm0) VOLUME_32x16 (%%xmm3, %%xmm2) + SWAP_16_2 (%%xmm0, %%xmm2) " movq %%xmm0, (%0) \n\t" " movq %%xmm2, 8(%0) \n\t" " add $16, %0 \n\t" @@ -198,29 +302,6 @@ pa_volume_s16ne_sse (int16_t *samples, int32_t *volumes, unsigned channels, unsi } #if 0 -static void -pa_volume_s16re_sse (int16_t *samples, int32_t *volumes, unsigned channels, unsigned length) -{ - unsigned channel; - - length /= sizeof (int16_t); - - for (channel = 0; length; length--) { - int32_t t, hi, lo; - - hi = volumes[channel] >> 16; - lo = volumes[channel] & 0xFFFF; - - t = (int32_t) PA_INT16_SWAP(*samples); - t = ((t * lo) >> 16) + (t * hi); - t = PA_CLAMP_UNLIKELY(t, -0x8000, 0x7FFF); - *samples++ = PA_INT16_SWAP((int16_t) t); - - if (PA_UNLIKELY(++channel >= channels)) - channel = 0; - } -} - static void pa_volume_float32ne_sse (float *samples, float *volumes, unsigned channels, unsigned length) { @@ -378,24 +459,25 @@ pa_volume_s24_32re_sse (uint32_t *samples, int32_t *volumes, unsigned channels, } #endif -#undef RUN_TEST +#define RUN_TEST #ifdef RUN_TEST #define CHANNELS 2 #define SAMPLES 1021 #define TIMES 1000 +#define PADDING 16 static void run_test (void) { int16_t samples[SAMPLES]; int16_t samples_ref[SAMPLES]; int16_t samples_orig[SAMPLES]; - int32_t volumes[CHANNELS + 16]; + int32_t volumes[CHANNELS + PADDING]; int i, j, padding; pa_do_volume_func_t func; - func = pa_get_volume_func (PA_SAMPLE_S16NE); + func = pa_get_volume_func (PA_SAMPLE_S16RE); - printf ("checking %d\n", sizeof (samples)); + printf ("checking SSE %d\n", sizeof (samples)); for (j = 0; j < TIMES; j++) { pa_random (samples, sizeof (samples)); @@ -403,12 +485,11 @@ static void run_test (void) { memcpy (samples_orig, samples, sizeof (samples)); for (i = 0; i < CHANNELS; i++) - volumes[i] = rand() >> 15; - - for (padding = 0; padding < 16; padding++, i++) + volumes[i] = rand() >> 1; + for (padding = 0; padding < PADDING; padding++, i++) volumes[i] = volumes[padding]; - pa_volume_s16ne_sse (samples, volumes, CHANNELS, SAMPLES * sizeof (int16_t)); + pa_volume_s16re_sse (samples, volumes, CHANNELS, SAMPLES * sizeof (int16_t)); func (samples_ref, volumes, CHANNELS, SAMPLES * sizeof (int16_t)); for (i = 0; i < SAMPLES; i++) { @@ -416,11 +497,6 @@ static void run_test (void) { printf ("%d: %04x != %04x (%04x * %04x)\n", i, samples[i], samples_ref[i], samples_orig[i], volumes[i % CHANNELS]); } -#if 0 - else - printf ("%d: %04x == %04x (%04x * %04x)\n", i, samples[i], samples_ref[i], - samples_orig[i], volumes[i % CHANNELS]); -#endif } } } @@ -434,4 +510,5 @@ void pa_volume_func_init_sse (void) { #endif pa_set_volume_func (PA_SAMPLE_S16NE, (pa_do_volume_func_t) pa_volume_s16ne_sse); + pa_set_volume_func (PA_SAMPLE_S16RE, (pa_do_volume_func_t) pa_volume_s16re_sse); } -- cgit