CLPF: Replace v128_shr_n_s8 with v128_cmplt_s8 for sign extraction
On x86 there is no direct v128_shr_n_s8 equivalent, so
v128_cmplt_s8(a, v128_zero()) is much better than v128_shr_n_s8(a, 7).
Should have no impact on NEON.
Also replace v256_from_v128(v128_from_v64(a, b), v128_from_v64(c, d))
with v256_from_v64(a, b, c, d).
Change-Id: I711e3cb250689089d7b5336a294e9d6bdd998445
diff --git a/av1/common/clpf_simd.h b/av1/common/clpf_simd.h
index 6ca3324..97fe87c 100644
--- a/av1/common/clpf_simd.h
+++ b/av1/common/clpf_simd.h
@@ -18,7 +18,7 @@
unsigned int adjdamp) {
const v256 diff16 = v256_sub_16(a, b);
v128 diff = v128_pack_s16_s8(v256_high_v128(diff16), v256_low_v128(diff16));
- const v128 sign = v128_shr_n_s8(diff, 7);
+ const v128 sign = v128_cmplt_s8(diff, v128_zero());
diff = v128_abs_s8(diff);
return v128_xor(
v128_add_8(sign,
@@ -115,34 +115,27 @@
const v64 l5 = v64_load_aligned(src + 3 * sstride);
const v64 l6 = v64_load_aligned(src + 4 * sstride);
const v64 l7 = v64_load_aligned(src + 5 * sstride);
- const v256 a = v256_from_v128(v128_from_v64(l0, l1), v128_from_v64(l2, l3));
- const v256 b = v256_from_v128(v128_from_v64(l1, l2), v128_from_v64(l3, l4));
- const v256 g = v256_from_v128(v128_from_v64(l3, l4), v128_from_v64(l5, l6));
- const v256 h = v256_from_v128(v128_from_v64(l4, l5), v128_from_v64(l6, l7));
- const v256 c = v256_from_v128(
- v128_from_v64(v64_load_unaligned(src - 2),
- v64_load_unaligned(src + sstride - 2)),
- v128_from_v64(v64_load_unaligned(src + 2 * sstride - 2),
- v64_load_unaligned(src + 3 * sstride - 2)));
- const v256 d = v256_from_v128(
- v128_from_v64(v64_load_unaligned(src - 1),
- v64_load_unaligned(src + sstride - 1)),
- v128_from_v64(v64_load_unaligned(src + 2 * sstride - 1),
- v64_load_unaligned(src + 3 * sstride - 1)));
- const v256 e = v256_from_v128(
- v128_from_v64(v64_load_unaligned(src + 1),
- v64_load_unaligned(src + sstride + 1)),
- v128_from_v64(v64_load_unaligned(src + 2 * sstride + 1),
- v64_load_unaligned(src + 3 * sstride + 1)));
- const v256 f = v256_from_v128(
- v128_from_v64(v64_load_unaligned(src + 2),
- v64_load_unaligned(src + sstride + 2)),
- v128_from_v64(v64_load_unaligned(src + 2 * sstride + 2),
- v64_load_unaligned(src + 3 * sstride + 2)));
-
const v128 o =
- calc_delta(v256_from_v128(v128_from_v64(l2, l3), v128_from_v64(l4, l5)),
- a, b, c, d, e, f, g, h, strength, adjdamp);
+ calc_delta(v256_from_v64(l2, l3, l4, l5), v256_from_v64(l0, l1, l2, l3),
+ v256_from_v64(l1, l2, l3, l4),
+ v256_from_v64(v64_load_unaligned(src - 2),
+ v64_load_unaligned(src + sstride - 2),
+ v64_load_unaligned(src + 2 * sstride - 2),
+ v64_load_unaligned(src + 3 * sstride - 2)),
+ v256_from_v64(v64_load_unaligned(src - 1),
+ v64_load_unaligned(src + sstride - 1),
+ v64_load_unaligned(src + 2 * sstride - 1),
+ v64_load_unaligned(src + 3 * sstride - 1)),
+ v256_from_v64(v64_load_unaligned(src + 1),
+ v64_load_unaligned(src + sstride + 1),
+ v64_load_unaligned(src + 2 * sstride + 1),
+ v64_load_unaligned(src + 3 * sstride + 1)),
+ v256_from_v64(v64_load_unaligned(src + 2),
+ v64_load_unaligned(src + sstride + 2),
+ v64_load_unaligned(src + 2 * sstride + 2),
+ v64_load_unaligned(src + 3 * sstride + 2)),
+ v256_from_v64(l3, l4, l5, l6), v256_from_v64(l4, l5, l6, l7),
+ strength, adjdamp);
u32_store_aligned(dst, v128_low_u32(v128_shr_n_byte(o, 12)));
u32_store_aligned(dst + dstride, v128_low_u32(v128_shr_n_byte(o, 8)));
@@ -188,34 +181,28 @@
int y;
for (y = 0; y < sizey; y += 4) {
- const v64 l0 = v64_load_aligned(src);
- const v64 l1 = v64_load_aligned(src + sstride);
- const v64 l2 = v64_load_aligned(src + 2 * sstride);
- const v64 l3 = v64_load_aligned(src + 3 * sstride);
- const v256 a = v256_from_v128(
- v128_from_v64(v64_load_unaligned(src - 2),
- v64_load_unaligned(src + sstride - 2)),
- v128_from_v64(v64_load_unaligned(src + 2 * sstride - 2),
- v64_load_unaligned(src + 3 * sstride - 2)));
- const v256 b = v256_from_v128(
- v128_from_v64(v64_load_unaligned(src - 1),
- v64_load_unaligned(src + sstride - 1)),
- v128_from_v64(v64_load_unaligned(src + 2 * sstride - 1),
- v64_load_unaligned(src + 3 * sstride - 1)));
- const v256 c = v256_from_v128(
- v128_from_v64(v64_load_unaligned(src + 1),
- v64_load_unaligned(src + sstride + 1)),
- v128_from_v64(v64_load_unaligned(src + 2 * sstride + 1),
- v64_load_unaligned(src + 3 * sstride + 1)));
- const v256 d = v256_from_v128(
- v128_from_v64(v64_load_unaligned(src + 2),
- v64_load_unaligned(src + sstride + 2)),
- v128_from_v64(v64_load_unaligned(src + 2 * sstride + 2),
- v64_load_unaligned(src + 3 * sstride + 2)));
+ const v256 a = v256_from_v64(v64_load_unaligned(src - 2),
+ v64_load_unaligned(src + sstride - 2),
+ v64_load_unaligned(src + 2 * sstride - 2),
+ v64_load_unaligned(src + 3 * sstride - 2));
+ const v256 b = v256_from_v64(v64_load_unaligned(src - 1),
+ v64_load_unaligned(src + sstride - 1),
+ v64_load_unaligned(src + 2 * sstride - 1),
+ v64_load_unaligned(src + 3 * sstride - 1));
+ const v256 c = v256_from_v64(v64_load_unaligned(src + 1),
+ v64_load_unaligned(src + sstride + 1),
+ v64_load_unaligned(src + 2 * sstride + 1),
+ v64_load_unaligned(src + 3 * sstride + 1));
+ const v256 d = v256_from_v64(v64_load_unaligned(src + 2),
+ v64_load_unaligned(src + sstride + 2),
+ v64_load_unaligned(src + 2 * sstride + 2),
+ v64_load_unaligned(src + 3 * sstride + 2));
const v128 o = calc_hdelta(
- v256_from_v128(v128_from_v64(l0, l1), v128_from_v64(l2, l3)), a, b, c,
- d, strength, adjdamp);
+ v256_from_v64(v64_load_aligned(src), v64_load_aligned(src + sstride),
+ v64_load_aligned(src + 2 * sstride),
+ v64_load_aligned(src + 3 * sstride)),
+ a, b, c, d, strength, adjdamp);
u32_store_aligned(dst, v128_low_u32(v128_shr_n_byte(o, 12)));
u32_store_aligned(dst + dstride, v128_low_u32(v128_shr_n_byte(o, 8)));
