Commit a67ae670 authored by Martin Storsjö's avatar Martin Storsjö

arm: vp9: Add NEON itxfm routines

This work is sponsored by, and copyright, Google.

For the transforms up to 8x8, we can fit all the data (including
temporaries) in registers and just do a straightforward transform
of all the data. For 16x16, we do a transform of 4x16 pixels in
4 slices, using a temporary buffer. For 32x32, we transform 4x32
pixels at a time, in two steps of 4x16 pixels each.

Examples of relative speedup compared to the C version, from checkasm:
                         Cortex       A7     A8     A9    A53
vp9_inv_adst_adst_4x4_add_neon:     3.39   5.83   4.17   4.01
vp9_inv_adst_adst_8x8_add_neon:     3.79   4.86   4.23   3.98
vp9_inv_adst_adst_16x16_add_neon:   3.33   4.36   4.11   4.16
vp9_inv_dct_dct_4x4_add_neon:       4.06   6.16   4.59   4.46
vp9_inv_dct_dct_8x8_add_neon:       4.61   6.01   4.98   4.86
vp9_inv_dct_dct_16x16_add_neon:     3.35   3.44   3.36   3.79
vp9_inv_dct_dct_32x32_add_neon:     3.89   3.50   3.79   4.42
vp9_inv_wht_wht_4x4_add_neon:       3.22   5.13   3.53   3.77

Thus, the speedup vs C code is around 3-6x.

This is mostly marginally faster than the corresponding routines
in libvpx on most cores, tested with their 32x32 idct (compared to
vpx_idct32x32_1024_add_neon). These numbers are slightly in libvpx's
favour since their version doesn't clear the input buffer like ours
do (although the effect of that on the total runtime probably is
negligible.)

                           Cortex       A7       A8       A9      A53
vp9_inv_dct_dct_32x32_add_neon:    18436.8  16874.1  14235.1  11988.9
libvpx vpx_idct32x32_1024_add_neon 20789.0  13344.3  15049.9  13030.5

Only on the Cortex A8, the libvpx function is faster. On the other cores,
ours is slightly faster even though ours has got source block clearing
integrated.
Signed-off-by: 's avatarMartin Storsjö <martin@martin.st>
parent 0b37cd09
......@@ -139,4 +139,5 @@ NEON-OBJS-$(CONFIG_RV40_DECODER) += arm/rv34dsp_neon.o \
arm/rv40dsp_neon.o
NEON-OBJS-$(CONFIG_VORBIS_DECODER) += arm/vorbisdsp_neon.o
NEON-OBJS-$(CONFIG_VP6_DECODER) += arm/vp6dsp_neon.o
NEON-OBJS-$(CONFIG_VP9_DECODER) += arm/vp9mc_neon.o
NEON-OBJS-$(CONFIG_VP9_DECODER) += arm/vp9itxfm_neon.o \
arm/vp9mc_neon.o
......@@ -94,7 +94,7 @@ define_8tap_2d_funcs(8)
define_8tap_2d_funcs(4)
av_cold void ff_vp9dsp_init_arm(VP9DSPContext *dsp)
static av_cold void vp9dsp_mc_init_arm(VP9DSPContext *dsp)
{
int cpu_flags = av_get_cpu_flags();
......@@ -138,3 +138,52 @@ av_cold void ff_vp9dsp_init_arm(VP9DSPContext *dsp)
init_mc_funcs_dirs(4, 4);
}
}
#define define_itxfm(type_a, type_b, sz) \
void ff_vp9_##type_a##_##type_b##_##sz##x##sz##_add_neon(uint8_t *_dst, \
ptrdiff_t stride, \
int16_t *_block, int eob)
#define define_itxfm_funcs(sz) \
define_itxfm(idct, idct, sz); \
define_itxfm(iadst, idct, sz); \
define_itxfm(idct, iadst, sz); \
define_itxfm(iadst, iadst, sz)
define_itxfm_funcs(4);
define_itxfm_funcs(8);
define_itxfm_funcs(16);
define_itxfm(idct, idct, 32);
define_itxfm(iwht, iwht, 4);
static av_cold void vp9dsp_itxfm_init_arm(VP9DSPContext *dsp)
{
int cpu_flags = av_get_cpu_flags();
if (have_neon(cpu_flags)) {
#define init_itxfm(tx, sz) \
dsp->itxfm_add[tx][DCT_DCT] = ff_vp9_idct_idct_##sz##_add_neon; \
dsp->itxfm_add[tx][DCT_ADST] = ff_vp9_iadst_idct_##sz##_add_neon; \
dsp->itxfm_add[tx][ADST_DCT] = ff_vp9_idct_iadst_##sz##_add_neon; \
dsp->itxfm_add[tx][ADST_ADST] = ff_vp9_iadst_iadst_##sz##_add_neon
#define init_idct(tx, nm) \
dsp->itxfm_add[tx][DCT_DCT] = \
dsp->itxfm_add[tx][ADST_DCT] = \
dsp->itxfm_add[tx][DCT_ADST] = \
dsp->itxfm_add[tx][ADST_ADST] = ff_vp9_##nm##_add_neon
init_itxfm(TX_4X4, 4x4);
init_itxfm(TX_8X8, 8x8);
init_itxfm(TX_16X16, 16x16);
init_idct(TX_32X32, idct_idct_32x32);
init_idct(4, iwht_iwht_4x4);
}
}
av_cold void ff_vp9dsp_init_arm(VP9DSPContext *dsp)
{
vp9dsp_mc_init_arm(dsp);
vp9dsp_itxfm_init_arm(dsp);
}
This diff is collapsed.
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