wrapper/FFmpeg
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

lavc/takdsp: R-V V decorrelate_sf

Browse source 
decorrelate_sf_c:      259.2
decorrelate_sf_rvv_i32: 45.5
This commit is contained in:
Rémi Denis-Courmont 2023-12-22 18:12:57 +02:00
parent 34a47b97de
commit 278b4b60d6
2 changed files with 23 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

2
libavcodec/riscv/takdsp_init.c
View file

@ -28,6 +28,7 @@
void ff_decorrelate_ls_rvv(const int32_t *p1, int32_t *p2, int length); void ff_decorrelate_ls_rvv(const int32_t *p1, int32_t *p2, int length);
void ff_decorrelate_sr_rvv(int32_t *p1, const int32_t *p2, int length); void ff_decorrelate_sr_rvv(int32_t *p1, const int32_t *p2, int length);
void ff_decorrelate_sm_rvv(int32_t *p1, int32_t *p2, int length); void ff_decorrelate_sm_rvv(int32_t *p1, int32_t *p2, int length);
void ff_decorrelate_sf_rvv(int32_t *p1, const int32_t *p2, int len, int, int);
av_cold void ff_takdsp_init_riscv(TAKDSPContext *dsp) av_cold void ff_takdsp_init_riscv(TAKDSPContext *dsp)
{ {
@ -38,6 +39,7 @@ av_cold void ff_takdsp_init_riscv(TAKDSPContext *dsp)
dsp->decorrelate_ls = ff_decorrelate_ls_rvv; dsp->decorrelate_ls = ff_decorrelate_ls_rvv;
dsp->decorrelate_sr = ff_decorrelate_sr_rvv; dsp->decorrelate_sr = ff_decorrelate_sr_rvv;
dsp->decorrelate_sm = ff_decorrelate_sm_rvv; dsp->decorrelate_sm = ff_decorrelate_sm_rvv;
dsp->decorrelate_sf = ff_decorrelate_sf_rvv;
} }
#endif #endif
} }

21
libavcodec/riscv/takdsp_rvv.S
View file

@ -1,5 +1,6 @@
/* /*
* Copyright (c) 2023 Institue of Software Chinese Academy of Sciences (ISCAS). * Copyright (c) 2023 Institue of Software Chinese Academy of Sciences (ISCAS).
* Copyright (c) 2023 Rémi Denis-Courmont
* *
* This file is part of FFmpeg. * This file is part of FFmpeg.
* *
@ -65,3 +66,23 @@ func ff_decorrelate_sm_rvv, zve32x
ret ret
endfunc endfunc
func ff_decorrelate_sf_rvv, zve32x
csrwi vxrm, 0
1:
vsetvli t0, a2, e32, m8, ta, ma
vle32.v v8, (a1)
sub a2, a2, t0
vsra.vx v8, v8, a3
sh2add a1, t0, a1
vle32.v v0, (a0)
vmul.vx v8, v8, a4
vssra.vi v8, v8, 8
vsll.vx v8, v8, a3
vsub.vv v0, v8, v0
vse32.v v0, (a0)
sh2add a0, t0, a0
bnez a2, 1b
ret
endfunc