virtualenv/FFmpeg
1
0
Fork 0
mirror of https://github.com/FFmpeg/FFmpeg.git synced 2025-01-24 13:56:33 +02:00
Code Issues Releases Activity

lavc/vp8dsp: use saturating add/sub for R-V V DC add

Browse Source 
T-Head C908 (cycles):
vp7_idct_dc_add_c:          108.5
vp7_idct_dc_add_rvv_i32:     56.2 (before)
vp7_idct_dc_add_rvv_i32:     47.2 (after)
vp8_idct_dc_add_c:           96.2
vp8_idct_dc_add_rvv_i32:     43.0 (before)
vp8_idct_dc_add_rvv_i32:     34.0 (after)
This commit is contained in:
Rémi Denis-Courmont 2024-07-25 17:40:26 +03:00
parent bbfc0ac9ca
commit 9b4655c3a1
1 changed files with 12 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
libavcodec/riscv/vp8dsp_rvv.S
View File

@ -172,12 +172,18 @@ func ff_vp78_idct_dc_add_rvv, zve32x
vsetivli zero, 4, e8, mf4, ta, ma vsetivli zero, 4, e8, mf4, ta, ma
sh zero, (a1) sh zero, (a1)
vlse32.v v8, (a0), a2 vlse32.v v8, (a0), a2
vsetivli zero, 16, e16, m2, ta, ma vsetivli zero, 16, e8, m1, ta, ma
vzext.vf2 v16, v8 bgez a3, 1f
vadd.vx v16, v16, a3
vmax.vx v16, v16, zero # block[0] < 0
vsetvli zero, zero, e8, m1, ta, ma neg a3, a3
vnclipu.wi v8, v16, 0 vssubu.vx v8, v8, a3
vsetivli zero, 4, e8, mf4, ta, ma
vsse32.v v8, (a0), a2
ret
1: # block[0] >= 0
vsaddu.vx v8, v8, a3
vsetivli zero, 4, e8, mf4, ta, ma vsetivli zero, 4, e8, mf4, ta, ma
vsse32.v v8, (a0), a2 vsse32.v v8, (a0), a2
ret ret