diff options
| author | Rémi Denis-Courmont <remi@remlab.net> | 2023-07-13 23:16:03 +0300 |
|---|---|---|
| committer | Rémi Denis-Courmont <remi@remlab.net> | 2023-07-16 23:24:00 +0300 |
| commit | c541ecf0dc38b6da3ba3d290b0db7d147775931f (patch) | |
| tree | fd3a6ef32d653a4155e28b7ef6cc43ef97767049 | |
| parent | a28aa0475db99cc66c37c98fd5330e313817709c (diff) | |
| download | ffmpeg-c541ecf0dc38b6da3ba3d290b0db7d147775931f.tar.gz | |
lavc/alacdsp: unroll RISC-V V loops
This increases the group multiplier as per T-Head C910 benchmarks:
alac_append_extra_bits_mono_c: 803.0
alac_append_extra_bits_stereo_c: 1604.2
alac_decorrelate_stereo_c: 1077.5
LMUL=1
alac_append_extra_bits_mono_rvv_i32: 418.2
alac_append_extra_bits_stereo_rvv_i32: 693.2
alac_decorrelate_stereo_rvv_i32: 673.5
LMUL=2
alac_append_extra_bits_mono_rvv_i32: 382.2
alac_append_extra_bits_stereo_rvv_i32: 648.2
alac_decorrelate_stereo_rvv_i32: 542.7
LMUL=4
alac_append_extra_bits_mono_rvv_i32: 241.5
alac_append_extra_bits_stereo_rvv_i32: 512.7
alac_decorrelate_stereo_rvv_i32: 364.2
LMUL=8
alac_append_extra_bits_mono_rvv_i32: 239.7
alac_append_extra_bits_stereo_rvv_i32: 497.2
alac_decorrelate_stereo_rvv_i32: 426.7
| -rw-r--r-- | libavcodec/riscv/alacdsp_rvv.S | 6 |
1 files changed, 3 insertions, 3 deletions
diff --git a/libavcodec/riscv/alacdsp_rvv.S b/libavcodec/riscv/alacdsp_rvv.S index 8fbe3fbe77..8efb04e0c8 100644 --- a/libavcodec/riscv/alacdsp_rvv.S +++ b/libavcodec/riscv/alacdsp_rvv.S @@ -25,7 +25,7 @@ func ff_alac_decorrelate_stereo_rvv, zve32x ld a4, 8(a0) ld a0, 0(a0) 1: - vsetvli t0, a1, e32, m1, ta, ma + vsetvli t0, a1, e32, m4, ta, ma vle32.v v24, (a4) sub a1, a1, t0 vle32.v v16, (a0) @@ -47,7 +47,7 @@ func ff_alac_append_extra_bits_mono_rvv, zve32x ld a0, (a0) ld a1, (a1) 1: - vsetvli t0, a4, e32, m1, ta, ma + vsetvli t0, a4, e32, m8, ta, ma vle32.v v16, (a0) sub a4, a4, t0 vle32.v v24, (a1) @@ -67,7 +67,7 @@ func ff_alac_append_extra_bits_stereo_rvv, zve32x ld a7, 8(a1) ld a1, (a1) 1: - vsetvli t0, a4, e32, m1, ta, ma + vsetvli t0, a4, e32, m8, ta, ma vle32.v v16, (a0) sub a4, a4, t0 vle32.v v0, (a6) |