mirrors/pipewire
1
0
Fork 0
mirror of https://gitlab.freedesktop.org/pipewire/pipewire.git synced 2025-11-02 09:01:50 -05:00
Code Issues Projects Releases Packages Wiki Activity Actions

fmt-ops: add f32 -> s16 see optimisations

Browse source
This commit is contained in:
Wim Taymans 2018-10-29 09:21:33 +00:00
parent 83517e22ac
commit 9f61ef96ff
2 changed files with 91 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

88
spa/plugins/audioconvert/fmt-ops-sse.c
View file

@ -273,3 +273,91 @@ conv_f32d_to_s32_sse(void *data, int n_dst, void *dst[n_dst], int n_src, const v
for(; i < n_src; i++)
conv_f32d_to_s32_1_sse(data, &d[i], n_src, &src[i], n_bytes);
}
static void
conv_f32d_to_s16_1_sse(void *data, void *dst, int n_src, const void *src[n_src], int n_bytes)
{
const float **s = (const float **) src;
const float *s0 = s[0];
int16_t *d = dst;
int n, n_samples, unrolled;
__m128 in[1];
__m128i out[4];
__m128i scale = _mm_set1_epi32(15 << 23);
n_samples = n_bytes / sizeof(float);
unrolled = n_samples / 4;
n_samples = n_samples & 3;
for(n = 0; unrolled--; n += 4) {
in[0] = _mm_loadu_ps(&s0[n]);
in[0] = (__m128)_mm_adds_epi16((__m128i)in[0], scale);
out[0] = _mm_cvtps_epi32(in[0]);
out[0] = _mm_packs_epi32(out[0], out[0]);
d[0*n_src] = _mm_extract_pi16(*(__m64*)out, 0);
d[1*n_src] = _mm_extract_pi16(*(__m64*)out, 1);
d[2*n_src] = _mm_extract_pi16(*(__m64*)out, 2);
d[3*n_src] = _mm_extract_pi16(*(__m64*)out, 3);
d += 4*n_src;
}
for(; n_samples--; n++) {
in[0] = _mm_load_ss(&s0[n]);
in[0] = (__m128)_mm_adds_epi16((__m128i)in[0], scale);
out[0] = _mm_cvttps_epi32(in[0]);
out[0] = _mm_packs_epi32(out[0], out[0]);
*d = _mm_extract_pi16(*(__m64*)out, 0);
d += n_src;
}
}
static void
conv_f32d_to_s16_2_sse(void *data, void *dst, int n_src, const void *src[n_src], int n_bytes)
{
const float **s = (const float **) src;
const float *s0 = s[0], *s1 = s[1];
int16_t *d = dst;
int n, n_samples, unrolled;
__m128 in[2], t[2];
__m128i out[4];
__m128i scale = _mm_set1_epi32(15 << 23);
n_samples = n_bytes / sizeof(float);
unrolled = n_samples / 4;
n_samples = n_samples & 3;
for(n = 0; unrolled--; n += 4) {
in[0] = _mm_loadu_ps(&s0[n]);
in[1] = _mm_loadu_ps(&s1[n]);
t[0] = _mm_unpacklo_ps(in[0], in[1]);
t[1] = _mm_unpackhi_ps(in[0], in[1]);
in[0] = (__m128)_mm_adds_epi16((__m128i)t[0], scale);
in[1] = (__m128)_mm_adds_epi16((__m128i)t[1], scale);
out[0] = _mm_cvtps_epi32(in[0]);
out[1] = _mm_cvtps_epi32(in[1]);
out[0] = _mm_packs_epi32(out[0], out[1]);
_mm_storeu_si128((__m128i*)d, out[0]);
d += 4*n_src;
}
for(; n_samples--; n++) {
in[0] = _mm_set_ps(s0[n], s1[n], 0, 0);
in[0] = (__m128)_mm_adds_epi16((__m128i)in[0], scale);
out[0] = _mm_cvtps_epi32(in[0]);
out[0] = _mm_packs_epi32(out[0], out[0]);
_mm_store_ss((float*)d, (__m128)out[0]);
d += n_src;
}
}
static void
conv_f32d_to_s16_sse(void *data, int n_dst, void *dst[n_dst], int n_src, const void *src[n_src], int n_bytes)
{
int16_t *d = dst[0];
int i = 0;
for(; i + 1 < n_src; i += 2)
conv_f32d_to_s16_2_sse(data, &d[i], n_src, &src[i], n_bytes);
for(; i < n_src; i++)
conv_f32d_to_s16_1_sse(data, &d[i], n_src, &src[i], n_bytes);
}

3
spa/plugins/audioconvert/fmt-ops.c
View file

@ -715,6 +715,9 @@ static const struct conv_info {
{ SPA_AUDIO_FORMAT_F32, SPA_AUDIO_FORMAT_S16, 0, conv_f32_to_s16 },
{ SPA_AUDIO_FORMAT_F32, SPA_AUDIO_FORMAT_S16P, 0, conv_f32_to_s16d },
#if defined (__SSE2__)
{ SPA_AUDIO_FORMAT_F32P, SPA_AUDIO_FORMAT_S16, FEATURE_SSE, conv_f32d_to_s16_sse },
#endif
{ SPA_AUDIO_FORMAT_F32P, SPA_AUDIO_FORMAT_S16, 0, conv_f32d_to_s16 },
{ SPA_AUDIO_FORMAT_F32, SPA_AUDIO_FORMAT_S32, 0, conv_f32_to_s32 },