[FFmpeg-cvslog] lavc/rv34dsp: R-V V rv34_idct_dc_add

[sunyuechi]

ffmpeg | branch: master | sunyuechi <sunyuechi at iscas.ac.cn> | Wed Jan 31 19:04:11 2024 +0800| [89189dd9e7aa9c90d5104a768b3397dc804a9f42] | committer: Rémi Denis-Courmont

lavc/rv34dsp: R-V V rv34_idct_dc_add

C908:
rv34_idct_dc_add_c: 134.7
rv34_idct_dc_add_rvv_i32: 45.5

Signed-off-by: Rémi Denis-Courmont <remi at remlab.net>

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=89189dd9e7aa9c90d5104a768b3397dc804a9f42
---

 libavcodec/riscv/rv34dsp_init.c |  2 ++
 libavcodec/riscv/rv34dsp_rvv.S  | 20 ++++++++++++++++++++
 2 files changed, 22 insertions(+)

diff --git a/libavcodec/riscv/rv34dsp_init.c b/libavcodec/riscv/rv34dsp_init.c
index 852c8ad9a8..7dcadc7e43 100644
--- a/libavcodec/riscv/rv34dsp_init.c
+++ b/libavcodec/riscv/rv34dsp_init.c
@@ -26,6 +26,7 @@
 #include "libavcodec/rv34dsp.h"
 
 void ff_rv34_inv_transform_dc_rvv(int16_t *block);
+void ff_rv34_idct_dc_add_rvv(uint8_t *dst, ptrdiff_t stride, int dc);
 
 av_cold void ff_rv34dsp_init_riscv(RV34DSPContext *c)
 {
@@ -34,6 +35,7 @@ av_cold void ff_rv34dsp_init_riscv(RV34DSPContext *c)
 
     if (flags & AV_CPU_FLAG_RVV_I32 && ff_get_rv_vlenb() >= 16) {
         c->rv34_inv_transform_dc = ff_rv34_inv_transform_dc_rvv;
+        c->rv34_idct_dc_add = ff_rv34_idct_dc_add_rvv;
     }
 #endif
 }
diff --git a/libavcodec/riscv/rv34dsp_rvv.S b/libavcodec/riscv/rv34dsp_rvv.S
index 21c203c0c4..f1f6345012 100644
--- a/libavcodec/riscv/rv34dsp_rvv.S
+++ b/libavcodec/riscv/rv34dsp_rvv.S
@@ -31,3 +31,23 @@ func ff_rv34_inv_transform_dc_rvv, zve32x
 
         ret
 endfunc
+
+func ff_rv34_idct_dc_add_rvv, zve32x
+        vsetivli      zero, 4, e8, mf4, ta, ma
+        vlse32.v      v0, (a0), a1
+        li            t1, 169
+        mul           t1, t1, a2
+        li            a2, 255
+        addi          t1, t1, 512
+        srai          t1, t1, 10
+        vsetivli      zero, 4*4, e16, m2, ta, ma
+        vzext.vf2     v2, v0
+        vadd.vx       v2, v2, t1
+        vmax.vx       v2, v2, zero
+        vsetvli       zero, zero, e8, m1, ta, ma
+        vnclipu.wi    v0, v2, 0
+        vsetivli      zero, 4, e8, mf4, ta, ma
+        vsse32.v      v0, (a0), a1
+
+        ret
+endfunc