lavc/me_cmp: R-V V pix_abs_x2

C908:
pix_abs_0_1_c: 767.0
pix_abs_0_1_rvv_i32: 196.2
pix_abs_1_1_c: 388.0
pix_abs_1_1_rvv_i32: 185.2

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

libavcodec/riscv/me_cmp_init.c

@@ -30,6 +30,10 @@ int ff_pix_abs16_rvv(MpegEncContext *v, const uint8_t *pix1, const uint8_t *pix2
                               ptrdiff_t stride, int h);
 int ff_pix_abs8_rvv(MpegEncContext *v, const uint8_t *pix1, const uint8_t *pix2,
                              ptrdiff_t stride, int h);
+int ff_pix_abs16_x2_rvv(MpegEncContext *v, const uint8_t *pix1, const uint8_t *pix2,
+                         ptrdiff_t stride, int h);
+int ff_pix_abs8_x2_rvv(MpegEncContext *v, const uint8_t *pix1, const uint8_t *pix2,
+                         ptrdiff_t stride, int h);
 
 av_cold void ff_me_cmp_init_riscv(MECmpContext *c, AVCodecContext *avctx)
 {
@@ -41,6 +45,8 @@ av_cold void ff_me_cmp_init_riscv(MECmpContext *c, AVCodecContext *avctx)
         c->sad[0] = ff_pix_abs16_rvv;
         c->pix_abs[1][0] = ff_pix_abs8_rvv;
         c->sad[1] = ff_pix_abs8_rvv;
+        c->pix_abs[0][1] = ff_pix_abs16_x2_rvv;
+        c->pix_abs[1][1] = ff_pix_abs8_x2_rvv;
     }
 #endif
 }

libavcodec/riscv/me_cmp_rvv.S

@@ -65,3 +65,53 @@ func ff_pix_abs8_rvv, zve32x
 
         pix_abs_ret
 endfunc
+
+func ff_pix_abs16_x2_rvv, zve32x
+        csrwi           vxrm, 0
+        vsetivli        zero, 1, e32, m1, ta, ma
+        li              t5, 1
+        vmv.s.x         v0, zero
+1:
+        vsetivli        zero, 17, e8, m2, tu, ma
+        vle8.v          v12, (a2)
+        addi            a4, a4, -1
+        vslide1down.vx  v24, v12, t5
+        vsetivli        zero, 16, e8, m1, tu, ma
+        vle8.v          v4, (a1)
+        vaaddu.vv       v12, v12, v24
+        vwsubu.vv       v16, v4, v12
+        add             a1, a1, a3
+        vwsubu.vv       v20, v12, v4
+        vsetvli         zero, zero, e16, m2, tu, ma
+        vmax.vv         v16, v16, v20
+        add             a2, a2, a3
+        vwredsum.vs     v0, v16, v0
+        bnez            a4, 1b
+
+        pix_abs_ret
+endfunc
+
+func ff_pix_abs8_x2_rvv, zve32x
+        csrwi           vxrm, 0
+        vsetivli        zero, 1, e32, m1, ta, ma
+        li              t5, 1
+        vmv.s.x         v0, zero
+1:
+        vsetivli        zero, 9, e8, m1, tu, ma
+        vle8.v          v12, (a2)
+        addi            a4, a4, -1
+        vslide1down.vx  v24, v12, t5
+        vsetivli        zero, 8, e8, mf2, tu, ma
+        vle8.v          v4, (a1)
+        vaaddu.vv       v12, v12, v24
+        vwsubu.vv       v16, v4, v12
+        add             a1, a1, a3
+        vwsubu.vv       v20, v12, v4
+        vsetvli         zero, zero, e16, m1, tu, ma
+        vmax.vv         v16, v16, v20
+        add             a2, a2, a3
+        vwredsum.vs     v0, v16, v0
+        bnez            a4, 1b
+
+        pix_abs_ret
+endfunc