RepoMirrors
/
ffmpeg
mirror of https://git.ffmpeg.org/ffmpeg.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

lavc/lpc: optimise RVV vector type for compute_autocorr 

On SpacemiT X60 (with len == 4000):
autocorr_10_c:       2303.7
autocorr_10_rvv_f64: 1411.5 (before)
autocorr_10_rvv_f64:  842.2 (after)
This commit is contained in:
Rémi Denis-Courmont 2024-05-26 08:25:16 +03:00
parent d82c503555
commit c53d42380d
2 changed files with 5 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
libavcodec/riscv/lpc_init.c
View File

@ -36,7 +36,8 @@ av_cold void ff_lpc_init_riscv(LPCContext *c)
if ((flags & AV_CPU_FLAG_RVV_F64) && (flags & AV_CPU_FLAG_RVB_ADDR)) {
c->lpc_apply_welch_window = ff_lpc_apply_welch_window_rvv;
if (ff_get_rv_vlenb() > c->max_order)
if ((flags & AV_CPU_FLAG_RVB_BASIC) &&
ff_get_rv_vlenb() > c->max_order)
c->lpc_compute_autocorr = ff_lpc_compute_autocorr_rvv;
}
#endif

5
libavcodec/riscv/lpc_rvv.S
View File

@ -86,10 +86,11 @@ func ff_lpc_apply_welch_window_rvv, zve64d
ret
endfunc
func ff_lpc_compute_autocorr_rvv, zve64d
func ff_lpc_compute_autocorr_rvv, zve64d, zbb
addi a2, a2, 1
vtype_vli t1, a2, t2, e64, ta, ma
li t0, 1
vsetvli zero, a2, e64, m8, ta, ma
vsetvl zero, a2, t1
fcvt.d.l ft0, t0
vle64.v v0, (a0)
sh3add a0, a2, a0 # data += lag