diff options
author | Justin Ruggles <justin.ruggles@gmail.com> | 2011-01-30 18:06:46 +0300 |
---|---|---|
committer | Michael Niedermayer <michaelni@gmx.at> | 2011-02-04 05:08:09 +0300 |
commit | fe2ff6d24745f0739bfde9061092c1268557310b (patch) | |
tree | 9cbcf8b2472dd7612dd84c8b6b237d9d02b4daf9 /libavcodec/ppc | |
parent | a35d782d28ef0497f2b65eb300c2e6a6028fc165 (diff) |
Separate format conversion DSP functions from DSPContext.
This will be beneficial for use with the audio conversion API without
requiring it to depend on all of dsputil.
Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit c73d99e672329c8f2df290736ffc474c360ac4ae)
Diffstat (limited to 'libavcodec/ppc')
-rw-r--r-- | libavcodec/ppc/Makefile | 1 | ||||
-rw-r--r-- | libavcodec/ppc/float_altivec.c | 112 | ||||
-rw-r--r-- | libavcodec/ppc/fmtconvert_altivec.c | 142 |
3 files changed, 143 insertions, 112 deletions
diff --git a/libavcodec/ppc/Makefile b/libavcodec/ppc/Makefile index 9b2358d49c..35ea0c38f8 100644 --- a/libavcodec/ppc/Makefile +++ b/libavcodec/ppc/Makefile @@ -21,6 +21,7 @@ ALTIVEC-OBJS-$(CONFIG_FFT) += ppc/fft_altivec.o \ OBJS-$(HAVE_ALTIVEC) += ppc/dsputil_altivec.o \ ppc/fdct_altivec.o \ ppc/float_altivec.o \ + ppc/fmtconvert_altivec.o \ ppc/gmc_altivec.o \ ppc/idct_altivec.o \ ppc/int_altivec.o \ diff --git a/libavcodec/ppc/float_altivec.c b/libavcodec/ppc/float_altivec.c index 60bae9a757..ba97cbfd3b 100644 --- a/libavcodec/ppc/float_altivec.c +++ b/libavcodec/ppc/float_altivec.c @@ -122,124 +122,12 @@ static void vector_fmul_window_altivec(float *dst, const float *src0, const floa } } -static void int32_to_float_fmul_scalar_altivec(float *dst, const int *src, float mul, int len) -{ - union { - vector float v; - float s[4]; - } mul_u; - int i; - vector float src1, src2, dst1, dst2, mul_v, zero; - - zero = (vector float)vec_splat_u32(0); - mul_u.s[0] = mul; - mul_v = vec_splat(mul_u.v, 0); - - for(i=0; i<len; i+=8) { - src1 = vec_ctf(vec_ld(0, src+i), 0); - src2 = vec_ctf(vec_ld(16, src+i), 0); - dst1 = vec_madd(src1, mul_v, zero); - dst2 = vec_madd(src2, mul_v, zero); - vec_st(dst1, 0, dst+i); - vec_st(dst2, 16, dst+i); - } -} - - -static vector signed short -float_to_int16_one_altivec(const float *src) -{ - vector float s0 = vec_ld(0, src); - vector float s1 = vec_ld(16, src); - vector signed int t0 = vec_cts(s0, 0); - vector signed int t1 = vec_cts(s1, 0); - return vec_packs(t0,t1); -} - -static void float_to_int16_altivec(int16_t *dst, const float *src, long len) -{ - int i; - vector signed short d0, d1, d; - vector unsigned char align; - if(((long)dst)&15) //FIXME - for(i=0; i<len-7; i+=8) { - d0 = vec_ld(0, dst+i); - d = float_to_int16_one_altivec(src+i); - d1 = vec_ld(15, dst+i); - d1 = vec_perm(d1, d0, vec_lvsl(0,dst+i)); - align = vec_lvsr(0, dst+i); - d0 = vec_perm(d1, d, align); - d1 = vec_perm(d, d1, align); - vec_st(d0, 0, dst+i); - vec_st(d1,15, dst+i); - } - else - for(i=0; i<len-7; i+=8) { - d = float_to_int16_one_altivec(src+i); - vec_st(d, 0, dst+i); - } -} - -static void -float_to_int16_interleave_altivec(int16_t *dst, const float **src, - long len, int channels) -{ - int i; - vector signed short d0, d1, d2, c0, c1, t0, t1; - vector unsigned char align; - if(channels == 1) - float_to_int16_altivec(dst, src[0], len); - else - if (channels == 2) { - if(((long)dst)&15) - for(i=0; i<len-7; i+=8) { - d0 = vec_ld(0, dst + i); - t0 = float_to_int16_one_altivec(src[0] + i); - d1 = vec_ld(31, dst + i); - t1 = float_to_int16_one_altivec(src[1] + i); - c0 = vec_mergeh(t0, t1); - c1 = vec_mergel(t0, t1); - d2 = vec_perm(d1, d0, vec_lvsl(0, dst + i)); - align = vec_lvsr(0, dst + i); - d0 = vec_perm(d2, c0, align); - d1 = vec_perm(c0, c1, align); - vec_st(d0, 0, dst + i); - d0 = vec_perm(c1, d2, align); - vec_st(d1, 15, dst + i); - vec_st(d0, 31, dst + i); - dst+=8; - } - else - for(i=0; i<len-7; i+=8) { - t0 = float_to_int16_one_altivec(src[0] + i); - t1 = float_to_int16_one_altivec(src[1] + i); - d0 = vec_mergeh(t0, t1); - d1 = vec_mergel(t0, t1); - vec_st(d0, 0, dst + i); - vec_st(d1, 16, dst + i); - dst+=8; - } - } else { - DECLARE_ALIGNED(16, int16_t, tmp)[len]; - int c, j; - for (c = 0; c < channels; c++) { - float_to_int16_altivec(tmp, src[c], len); - for (i = 0, j = c; i < len; i++, j+=channels) { - dst[j] = tmp[i]; - } - } - } -} - void float_init_altivec(DSPContext* c, AVCodecContext *avctx) { c->vector_fmul = vector_fmul_altivec; c->vector_fmul_reverse = vector_fmul_reverse_altivec; c->vector_fmul_add = vector_fmul_add_altivec; - c->int32_to_float_fmul_scalar = int32_to_float_fmul_scalar_altivec; if(!(avctx->flags & CODEC_FLAG_BITEXACT)) { c->vector_fmul_window = vector_fmul_window_altivec; - c->float_to_int16 = float_to_int16_altivec; - c->float_to_int16_interleave = float_to_int16_interleave_altivec; } } diff --git a/libavcodec/ppc/fmtconvert_altivec.c b/libavcodec/ppc/fmtconvert_altivec.c new file mode 100644 index 0000000000..e5287c96c1 --- /dev/null +++ b/libavcodec/ppc/fmtconvert_altivec.c @@ -0,0 +1,142 @@ +/* + * Copyright (c) 2006 Luca Barbato <lu_zero@gentoo.org> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "libavcodec/fmtconvert.h" + +#include "dsputil_altivec.h" +#include "util_altivec.h" + +static void int32_to_float_fmul_scalar_altivec(float *dst, const int *src, float mul, int len) +{ + union { + vector float v; + float s[4]; + } mul_u; + int i; + vector float src1, src2, dst1, dst2, mul_v, zero; + + zero = (vector float)vec_splat_u32(0); + mul_u.s[0] = mul; + mul_v = vec_splat(mul_u.v, 0); + + for(i=0; i<len; i+=8) { + src1 = vec_ctf(vec_ld(0, src+i), 0); + src2 = vec_ctf(vec_ld(16, src+i), 0); + dst1 = vec_madd(src1, mul_v, zero); + dst2 = vec_madd(src2, mul_v, zero); + vec_st(dst1, 0, dst+i); + vec_st(dst2, 16, dst+i); + } +} + + +static vector signed short +float_to_int16_one_altivec(const float *src) +{ + vector float s0 = vec_ld(0, src); + vector float s1 = vec_ld(16, src); + vector signed int t0 = vec_cts(s0, 0); + vector signed int t1 = vec_cts(s1, 0); + return vec_packs(t0,t1); +} + +static void float_to_int16_altivec(int16_t *dst, const float *src, long len) +{ + int i; + vector signed short d0, d1, d; + vector unsigned char align; + if(((long)dst)&15) //FIXME + for(i=0; i<len-7; i+=8) { + d0 = vec_ld(0, dst+i); + d = float_to_int16_one_altivec(src+i); + d1 = vec_ld(15, dst+i); + d1 = vec_perm(d1, d0, vec_lvsl(0,dst+i)); + align = vec_lvsr(0, dst+i); + d0 = vec_perm(d1, d, align); + d1 = vec_perm(d, d1, align); + vec_st(d0, 0, dst+i); + vec_st(d1,15, dst+i); + } + else + for(i=0; i<len-7; i+=8) { + d = float_to_int16_one_altivec(src+i); + vec_st(d, 0, dst+i); + } +} + +static void +float_to_int16_interleave_altivec(int16_t *dst, const float **src, + long len, int channels) +{ + int i; + vector signed short d0, d1, d2, c0, c1, t0, t1; + vector unsigned char align; + if(channels == 1) + float_to_int16_altivec(dst, src[0], len); + else + if (channels == 2) { + if(((long)dst)&15) + for(i=0; i<len-7; i+=8) { + d0 = vec_ld(0, dst + i); + t0 = float_to_int16_one_altivec(src[0] + i); + d1 = vec_ld(31, dst + i); + t1 = float_to_int16_one_altivec(src[1] + i); + c0 = vec_mergeh(t0, t1); + c1 = vec_mergel(t0, t1); + d2 = vec_perm(d1, d0, vec_lvsl(0, dst + i)); + align = vec_lvsr(0, dst + i); + d0 = vec_perm(d2, c0, align); + d1 = vec_perm(c0, c1, align); + vec_st(d0, 0, dst + i); + d0 = vec_perm(c1, d2, align); + vec_st(d1, 15, dst + i); + vec_st(d0, 31, dst + i); + dst+=8; + } + else + for(i=0; i<len-7; i+=8) { + t0 = float_to_int16_one_altivec(src[0] + i); + t1 = float_to_int16_one_altivec(src[1] + i); + d0 = vec_mergeh(t0, t1); + d1 = vec_mergel(t0, t1); + vec_st(d0, 0, dst + i); + vec_st(d1, 16, dst + i); + dst+=8; + } + } else { + DECLARE_ALIGNED(16, int16_t, tmp)[len]; + int c, j; + for (c = 0; c < channels; c++) { + float_to_int16_altivec(tmp, src[c], len); + for (i = 0, j = c; i < len; i++, j+=channels) { + dst[j] = tmp[i]; + } + } + } +} + +void ff_fmt_convert_init_ppc(FmtConvertContext *c, AVCodecContext *avctx) +{ + c->int32_to_float_fmul_scalar = int32_to_float_fmul_scalar_altivec; + if(!(avctx->flags & CODEC_FLAG_BITEXACT)) { + c->float_to_int16 = float_to_int16_altivec; + c->float_to_int16_interleave = float_to_int16_interleave_altivec; + } +} |