Refactor Neon compound convolution functions 6/7
Refactor the Armv8.0 MLA implementation of the horizontal pass of
av1_dist_wtd_convolve_2d_neon. This is mostly just cosmetic changes,
as well as copying the convolution helper functions into the .c file
from the convolve_neon.h header.
Change-Id: Idfe1f20577a75e216dc338872ff62d420eb4f6ac
diff --git a/av1/common/arm/jnt_convolve_neon.c b/av1/common/arm/jnt_convolve_neon.c
index 02f6370..53c909d 100644
--- a/av1/common/arm/jnt_convolve_neon.c
+++ b/av1/common/arm/jnt_convolve_neon.c
@@ -179,7 +179,7 @@
#if defined(__aarch64__) && defined(__ARM_FEATURE_MATMUL_INT8)
-static INLINE void dist_wtd_convolve_2d_horiz_neon(
+static INLINE void dist_wtd_convolve_2d_horiz_8tap_neon(
const uint8_t *src, int src_stride, int16_t *im_block, const int im_stride,
const int16x8_t x_filter_s16, const int im_h, int w) {
const int bd = 8;
@@ -257,7 +257,7 @@
#elif defined(__aarch64__) && defined(__ARM_FEATURE_DOTPROD)
-static INLINE void dist_wtd_convolve_2d_horiz_neon(
+static INLINE void dist_wtd_convolve_2d_horiz_8tap_neon(
const uint8_t *src, int src_stride, int16_t *im_block, const int im_stride,
const int16x8_t x_filter_s16, const int im_h, int w) {
const int bd = 8;
@@ -343,100 +343,129 @@
#else // !(defined(__aarch64__) && defined(__ARM_FEATURE_DOTPROD))
-static INLINE void dist_wtd_convolve_2d_horiz_neon(
+static INLINE int16x4_t convolve8_4_2d_h(const int16x4_t s0, const int16x4_t s1,
+ const int16x4_t s2, const int16x4_t s3,
+ const int16x4_t s4, const int16x4_t s5,
+ const int16x4_t s6, const int16x4_t s7,
+ const int16x8_t x_filter,
+ const int16x4_t horiz_const) {
+ const int16x4_t x_filter_0_3 = vget_low_s16(x_filter);
+ const int16x4_t x_filter_4_7 = vget_high_s16(x_filter);
+
+ int16x4_t sum = horiz_const;
+ sum = vmla_lane_s16(sum, s0, x_filter_0_3, 0);
+ sum = vmla_lane_s16(sum, s1, x_filter_0_3, 1);
+ sum = vmla_lane_s16(sum, s2, x_filter_0_3, 2);
+ sum = vmla_lane_s16(sum, s3, x_filter_0_3, 3);
+ sum = vmla_lane_s16(sum, s4, x_filter_4_7, 0);
+ sum = vmla_lane_s16(sum, s5, x_filter_4_7, 1);
+ sum = vmla_lane_s16(sum, s6, x_filter_4_7, 2);
+ sum = vmla_lane_s16(sum, s7, x_filter_4_7, 3);
+
+ // We halved the convolution filter values so -1 from the right shift.
+ return vshr_n_s16(sum, ROUND0_BITS - 1);
+}
+
+static INLINE int16x8_t convolve8_8_2d_h(const int16x8_t s0, const int16x8_t s1,
+ const int16x8_t s2, const int16x8_t s3,
+ const int16x8_t s4, const int16x8_t s5,
+ const int16x8_t s6, const int16x8_t s7,
+ const int16x8_t x_filter,
+ const int16x8_t horiz_const) {
+ const int16x4_t x_filter_0_3 = vget_low_s16(x_filter);
+ const int16x4_t x_filter_4_7 = vget_high_s16(x_filter);
+
+ int16x8_t sum = horiz_const;
+ sum = vmlaq_lane_s16(sum, s0, x_filter_0_3, 0);
+ sum = vmlaq_lane_s16(sum, s1, x_filter_0_3, 1);
+ sum = vmlaq_lane_s16(sum, s2, x_filter_0_3, 2);
+ sum = vmlaq_lane_s16(sum, s3, x_filter_0_3, 3);
+ sum = vmlaq_lane_s16(sum, s4, x_filter_4_7, 0);
+ sum = vmlaq_lane_s16(sum, s5, x_filter_4_7, 1);
+ sum = vmlaq_lane_s16(sum, s6, x_filter_4_7, 2);
+ sum = vmlaq_lane_s16(sum, s7, x_filter_4_7, 3);
+
+ // We halved the convolution filter values so -1 from the right shift.
+ return vshrq_n_s16(sum, ROUND0_BITS - 1);
+}
+
+static INLINE void dist_wtd_convolve_2d_horiz_8tap_neon(
const uint8_t *src, int src_stride, int16_t *im_block, const int im_stride,
const int16x8_t x_filter, const int im_h, int w) {
const int bd = 8;
- const uint8_t *s;
- int16_t *dst_ptr;
- int dst_stride;
- int width, height;
- dst_ptr = im_block;
- dst_stride = im_stride;
- height = im_h;
- width = w;
+ const uint8_t *src_ptr = src;
+ int16_t *dst_ptr = im_block;
+ int dst_stride = im_stride;
+ int height = im_h;
if (w == 4) {
int16x4_t s0, s1, s2, s3, s4, s5, s6, s7, d0;
- int16x8_t tt0;
uint8x8_t t0;
-
- // This shim of 1 << ((ROUND0_BITS - 1) - 1) enables us to use non-rounding
- // shifts - which are generally faster than rounding shifts on modern CPUs.
- // The outermost -1 is needed because we halved the filter values.
- const int16x4_t horiz_const = vdup_n_s16((1 << (bd + FILTER_BITS - 2)) +
- (1 << ((ROUND0_BITS - 1) - 1)));
-
#if defined(__aarch64__)
int16x4_t s8, s9, s10, d1, d2, d3;
- int16x8_t tt1, tt2, tt3;
uint8x8_t t1, t2, t3;
#endif // defined(__aarch64__)
- do {
- s = src;
- __builtin_prefetch(s + 0 * src_stride);
-#if defined(__aarch64__)
- __builtin_prefetch(s + 1 * src_stride);
- __builtin_prefetch(s + 2 * src_stride);
- __builtin_prefetch(s + 3 * src_stride);
- load_u8_8x4(s, src_stride, &t0, &t1, &t2, &t3);
+ // A shim of 1 << ((ROUND0_BITS - 1) - 1) enables us to use non-rounding
+ // shifts - which are generally faster than rounding shifts on modern CPUs.
+ // (The extra -1 is needed because we halved the filter values.)
+ const int16x4_t horiz_const = vdup_n_s16((1 << (bd + FILTER_BITS - 2)) +
+ (1 << ((ROUND0_BITS - 1) - 1)));
+ do {
+ __builtin_prefetch(src_ptr + 0 * src_stride);
+#if defined(__aarch64__)
+ __builtin_prefetch(src_ptr + 1 * src_stride);
+ __builtin_prefetch(src_ptr + 2 * src_stride);
+ __builtin_prefetch(src_ptr + 3 * src_stride);
+
+ load_u8_8x4(src_ptr, src_stride, &t0, &t1, &t2, &t3);
transpose_u8_8x4(&t0, &t1, &t2, &t3);
- tt0 = vreinterpretq_s16_u16(vmovl_u8(t0));
- tt1 = vreinterpretq_s16_u16(vmovl_u8(t1));
- tt2 = vreinterpretq_s16_u16(vmovl_u8(t2));
- tt3 = vreinterpretq_s16_u16(vmovl_u8(t3));
- s0 = vget_low_s16(tt0);
- s1 = vget_low_s16(tt1);
- s2 = vget_low_s16(tt2);
- s3 = vget_low_s16(tt3);
- s4 = vget_high_s16(tt0);
- s5 = vget_high_s16(tt1);
- s6 = vget_high_s16(tt2);
+
+ s0 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t0)));
+ s1 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t1)));
+ s2 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t2)));
+ s3 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t3)));
+ s4 = vget_high_s16(vreinterpretq_s16_u16(vmovl_u8(t0)));
+ s5 = vget_high_s16(vreinterpretq_s16_u16(vmovl_u8(t1)));
+ s6 = vget_high_s16(vreinterpretq_s16_u16(vmovl_u8(t2)));
+
__builtin_prefetch(dst_ptr + 0 * dst_stride);
__builtin_prefetch(dst_ptr + 1 * dst_stride);
__builtin_prefetch(dst_ptr + 2 * dst_stride);
__builtin_prefetch(dst_ptr + 3 * dst_stride);
- s += 7;
- load_u8_8x4(s, src_stride, &t0, &t1, &t2, &t3);
+ load_u8_8x4(src_ptr + 7, src_stride, &t0, &t1, &t2, &t3);
transpose_u8_8x4(&t0, &t1, &t2, &t3);
- tt0 = vreinterpretq_s16_u16(vmovl_u8(t0));
- tt1 = vreinterpretq_s16_u16(vmovl_u8(t1));
- tt2 = vreinterpretq_s16_u16(vmovl_u8(t2));
- tt3 = vreinterpretq_s16_u16(vmovl_u8(t3));
- s7 = vget_low_s16(tt0);
- s8 = vget_low_s16(tt1);
- s9 = vget_low_s16(tt2);
- s10 = vget_low_s16(tt3);
- d0 = convolve8_horiz_4x4_s16(s0, s1, s2, s3, s4, s5, s6, s7, x_filter,
- horiz_const);
- d1 = convolve8_horiz_4x4_s16(s1, s2, s3, s4, s5, s6, s7, s8, x_filter,
- horiz_const);
- d2 = convolve8_horiz_4x4_s16(s2, s3, s4, s5, s6, s7, s8, s9, x_filter,
- horiz_const);
- d3 = convolve8_horiz_4x4_s16(s3, s4, s5, s6, s7, s8, s9, s10, x_filter,
- horiz_const);
+ s7 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t0)));
+ s8 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t1)));
+ s9 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t2)));
+ s10 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t3)));
+
+ d0 = convolve8_4_2d_h(s0, s1, s2, s3, s4, s5, s6, s7, x_filter,
+ horiz_const);
+ d1 = convolve8_4_2d_h(s1, s2, s3, s4, s5, s6, s7, s8, x_filter,
+ horiz_const);
+ d2 = convolve8_4_2d_h(s2, s3, s4, s5, s6, s7, s8, s9, x_filter,
+ horiz_const);
+ d3 = convolve8_4_2d_h(s3, s4, s5, s6, s7, s8, s9, s10, x_filter,
+ horiz_const);
transpose_s16_4x4d(&d0, &d1, &d2, &d3);
-
store_s16_4x4(dst_ptr, dst_stride, d0, d1, d2, d3);
- src += 4 * src_stride;
+ src_ptr += 4 * src_stride;
dst_ptr += 4 * dst_stride;
height -= 4;
#else // !defined(__aarch64__)
- t0 = vld1_u8(s); // a0 a1 a2 a3 a4 a5 a6 a7
- tt0 = vreinterpretq_s16_u16(vmovl_u8(t0)); // a0 a1 a2 a3 a4 a5 a6 a7
- s0 = vget_low_s16(tt0); // a0 a1 a2 a3
- s4 = vget_high_s16(tt0); // a4 a5 a6 a7
- __builtin_prefetch(dst_ptr);
- s += 8;
- t0 = vld1_u8(s); // a8 a9 a10 a11
+ t0 = vld1_u8(src_ptr); // a0 a1 a2 a3 a4 a5 a6 a7
+ s0 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t0))); // a0 a1 a2 a3
+ s4 = vget_high_s16(vreinterpretq_s16_u16(vmovl_u8(t0))); // a4 a5 a6 a7
- // a8 a9 a10 a11
+ __builtin_prefetch(dst_ptr);
+
+ t0 = vld1_u8(src_ptr + 8); // a8 a9 a10 a11
s7 = vget_low_s16(vreinterpretq_s16_u16(vmovl_u8(t0)));
s1 = vext_s16(s0, s4, 1); // a1 a2 a3 a4
@@ -446,42 +475,47 @@
s6 = vext_s16(s4, s7, 2); // a6 a7 a8 a9
s7 = vext_s16(s4, s7, 3); // a7 a8 a9 a10
- d0 = convolve8_horiz_4x4_s16(s0, s1, s2, s3, s4, s5, s6, s7, x_filter,
- horiz_const);
-
+ d0 = convolve8_4_2d_h(s0, s1, s2, s3, s4, s5, s6, s7, x_filter,
+ horiz_const);
vst1_s16(dst_ptr, d0);
- src += src_stride;
+ src_ptr += src_stride;
dst_ptr += dst_stride;
- height -= 1;
+ height--;
#endif // defined(__aarch64__)
} while (height > 0);
} else {
- int16_t *d_tmp;
- int16x8_t s0, s1, s2, s3, s4, s5, s6, s7;
- int16x8_t res0;
+ int16x8_t s0, s1, s2, s3, s4, s5, s6, s7, s8, d0;
uint8x8_t t0;
+#if defined(__aarch64__)
+ int16x8_t s9, s10, s11, s12, s13, s14;
+ int16x8_t d1, d2, d3, d4, d5, d6, d7;
+ uint8x8_t t1, t2, t3, t4, t5, t6, t7;
+#endif // defined(__aarch64__)
- // This shim of 1 << ((ROUND0_BITS - 1) - 1) enables us to use non-rounding
+ // A shim of 1 << ((ROUND0_BITS - 1) - 1) enables us to use non-rounding
// shifts - which are generally faster than rounding shifts on modern CPUs.
- // The outermost -1 is needed because we halved the filter values.
+ // (The extra -1 is needed because we halved the filter values.)
const int16x8_t horiz_const = vdupq_n_s16((1 << (bd + FILTER_BITS - 2)) +
(1 << ((ROUND0_BITS - 1) - 1)));
do {
+ const uint8_t *s;
+ int16_t *d = dst_ptr;
+ int width = w;
+
#if defined(__aarch64__)
- uint8x8_t t1, t2, t3, t4, t5, t6, t7;
- int16x8_t s8, s9, s10, s11, s12, s13, s14;
- int16x8_t res1, res2, res3, res4, res5, res6, res7;
- __builtin_prefetch(src + 0 * src_stride);
- __builtin_prefetch(src + 1 * src_stride);
- __builtin_prefetch(src + 2 * src_stride);
- __builtin_prefetch(src + 3 * src_stride);
- __builtin_prefetch(src + 4 * src_stride);
- __builtin_prefetch(src + 5 * src_stride);
- __builtin_prefetch(src + 6 * src_stride);
- __builtin_prefetch(src + 7 * src_stride);
- load_u8_8x8(src, src_stride, &t0, &t1, &t2, &t3, &t4, &t5, &t6, &t7);
+ __builtin_prefetch(src_ptr + 0 * src_stride);
+ __builtin_prefetch(src_ptr + 1 * src_stride);
+ __builtin_prefetch(src_ptr + 2 * src_stride);
+ __builtin_prefetch(src_ptr + 3 * src_stride);
+ __builtin_prefetch(src_ptr + 4 * src_stride);
+ __builtin_prefetch(src_ptr + 5 * src_stride);
+ __builtin_prefetch(src_ptr + 6 * src_stride);
+ __builtin_prefetch(src_ptr + 7 * src_stride);
+
+ load_u8_8x8(src_ptr, src_stride, &t0, &t1, &t2, &t3, &t4, &t5, &t6, &t7);
transpose_u8_8x8(&t0, &t1, &t2, &t3, &t4, &t5, &t6, &t7);
+
s0 = vreinterpretq_s16_u16(vmovl_u8(t0));
s1 = vreinterpretq_s16_u16(vmovl_u8(t1));
s2 = vreinterpretq_s16_u16(vmovl_u8(t2));
@@ -490,9 +524,8 @@
s5 = vreinterpretq_s16_u16(vmovl_u8(t5));
s6 = vreinterpretq_s16_u16(vmovl_u8(t6));
- width = w;
- s = src + 7;
- d_tmp = dst_ptr;
+ s = src_ptr + 7;
+
__builtin_prefetch(dst_ptr + 0 * dst_stride);
__builtin_prefetch(dst_ptr + 1 * dst_stride);
__builtin_prefetch(dst_ptr + 2 * dst_stride);
@@ -505,6 +538,7 @@
do {
load_u8_8x8(s, src_stride, &t0, &t1, &t2, &t3, &t4, &t5, &t6, &t7);
transpose_u8_8x8(&t0, &t1, &t2, &t3, &t4, &t5, &t6, &t7);
+
s7 = vreinterpretq_s16_u16(vmovl_u8(t0));
s8 = vreinterpretq_s16_u16(vmovl_u8(t1));
s9 = vreinterpretq_s16_u16(vmovl_u8(t2));
@@ -514,28 +548,26 @@
s13 = vreinterpretq_s16_u16(vmovl_u8(t6));
s14 = vreinterpretq_s16_u16(vmovl_u8(t7));
- res0 = convolve8_horiz_8x8_s16(s0, s1, s2, s3, s4, s5, s6, s7, x_filter,
- horiz_const);
- res1 = convolve8_horiz_8x8_s16(s1, s2, s3, s4, s5, s6, s7, s8, x_filter,
- horiz_const);
- res2 = convolve8_horiz_8x8_s16(s2, s3, s4, s5, s6, s7, s8, s9, x_filter,
- horiz_const);
- res3 = convolve8_horiz_8x8_s16(s3, s4, s5, s6, s7, s8, s9, s10,
- x_filter, horiz_const);
- res4 = convolve8_horiz_8x8_s16(s4, s5, s6, s7, s8, s9, s10, s11,
- x_filter, horiz_const);
- res5 = convolve8_horiz_8x8_s16(s5, s6, s7, s8, s9, s10, s11, s12,
- x_filter, horiz_const);
- res6 = convolve8_horiz_8x8_s16(s6, s7, s8, s9, s10, s11, s12, s13,
- x_filter, horiz_const);
- res7 = convolve8_horiz_8x8_s16(s7, s8, s9, s10, s11, s12, s13, s14,
- x_filter, horiz_const);
+ d0 = convolve8_8_2d_h(s0, s1, s2, s3, s4, s5, s6, s7, x_filter,
+ horiz_const);
+ d1 = convolve8_8_2d_h(s1, s2, s3, s4, s5, s6, s7, s8, x_filter,
+ horiz_const);
+ d2 = convolve8_8_2d_h(s2, s3, s4, s5, s6, s7, s8, s9, x_filter,
+ horiz_const);
+ d3 = convolve8_8_2d_h(s3, s4, s5, s6, s7, s8, s9, s10, x_filter,
+ horiz_const);
+ d4 = convolve8_8_2d_h(s4, s5, s6, s7, s8, s9, s10, s11, x_filter,
+ horiz_const);
+ d5 = convolve8_8_2d_h(s5, s6, s7, s8, s9, s10, s11, s12, x_filter,
+ horiz_const);
+ d6 = convolve8_8_2d_h(s6, s7, s8, s9, s10, s11, s12, s13, x_filter,
+ horiz_const);
+ d7 = convolve8_8_2d_h(s7, s8, s9, s10, s11, s12, s13, s14, x_filter,
+ horiz_const);
- transpose_s16_8x8(&res0, &res1, &res2, &res3, &res4, &res5, &res6,
- &res7);
+ transpose_s16_8x8(&d0, &d1, &d2, &d3, &d4, &d5, &d6, &d7);
+ store_s16_8x8(d, dst_stride, d0, d1, d2, d3, d4, d5, d6, d7);
- store_s16_8x8(d_tmp, dst_stride, res0, res1, res2, res3, res4, res5,
- res6, res7);
s0 = s8;
s1 = s9;
s2 = s10;
@@ -544,47 +576,43 @@
s5 = s13;
s6 = s14;
s += 8;
- d_tmp += 8;
+ d += 8;
width -= 8;
} while (width > 0);
- src += 8 * src_stride;
+ src_ptr += 8 * src_stride;
dst_ptr += 8 * dst_stride;
height -= 8;
#else // !defined(__aarch64__)
- int16x8_t temp_0;
- t0 = vld1_u8(src);
+ t0 = vld1_u8(src_ptr);
s0 = vreinterpretq_s16_u16(vmovl_u8(t0)); // a0 a1 a2 a3 a4 a5 a6 a7
- width = w;
- s = src + 8;
- d_tmp = dst_ptr;
+ s = src_ptr + 8;
__builtin_prefetch(dst_ptr);
do {
t0 = vld1_u8(s); // a8 a9 a10 a11 a12 a13 a14 a15
- s7 = vreinterpretq_s16_u16(vmovl_u8(t0));
- temp_0 = s0;
- s0 = s7;
+ s8 = vreinterpretq_s16_u16(vmovl_u8(t0));
- s1 = vextq_s16(temp_0, s7, 1); // a1 a2 a3 a4 a5 a6 a7 a8
- s2 = vextq_s16(temp_0, s7, 2); // a2 a3 a4 a5 a6 a7 a8 a9
- s3 = vextq_s16(temp_0, s7, 3); // a3 a4 a5 a6 a7 a8 a9 a10
- s4 = vextq_s16(temp_0, s7, 4); // a4 a5 a6 a7 a8 a9 a10 a11
- s5 = vextq_s16(temp_0, s7, 5); // a5 a6 a7 a8 a9 a10 a11 a12
- s6 = vextq_s16(temp_0, s7, 6); // a6 a7 a8 a9 a10 a11 a12 a13
- s7 = vextq_s16(temp_0, s7, 7); // a7 a8 a9 a10 a11 a12 a13 a14
+ s1 = vextq_s16(s0, s8, 1); // a1 a2 a3 a4 a5 a6 a7 a8
+ s2 = vextq_s16(s0, s8, 2); // a2 a3 a4 a5 a6 a7 a8 a9
+ s3 = vextq_s16(s0, s8, 3); // a3 a4 a5 a6 a7 a8 a9 a10
+ s4 = vextq_s16(s0, s8, 4); // a4 a5 a6 a7 a8 a9 a10 a11
+ s5 = vextq_s16(s0, s8, 5); // a5 a6 a7 a8 a9 a10 a11 a12
+ s6 = vextq_s16(s0, s8, 6); // a6 a7 a8 a9 a10 a11 a12 a13
+ s7 = vextq_s16(s0, s8, 7); // a7 a8 a9 a10 a11 a12 a13 a14
- res0 = convolve8_horiz_8x8_s16(temp_0, s1, s2, s3, s4, s5, s6, s7,
- x_filter, horiz_const);
- vst1q_s16(d_tmp, res0);
+ d0 = convolve8_8_2d_h(s0, s1, s2, s3, s4, s5, s6, s7, x_filter,
+ horiz_const);
+ vst1q_s16(d, d0);
+ s0 = s8;
s += 8;
- d_tmp += 8;
+ d += 8;
width -= 8;
} while (width > 0);
- src += src_stride;
+ src_ptr += src_stride;
dst_ptr += dst_stride;
- height -= 1;
+ height--;
#endif // defined(__aarch64__)
} while (height > 0);
}
@@ -1068,8 +1096,8 @@
const InterpFilterParams *filter_params_y,
const int subpel_x_qn, const int subpel_y_qn,
ConvolveParams *conv_params) {
- assert(!(w % 4));
- assert(!(h % 4));
+ assert(w % 4 == 0);
+ assert(h % 4 == 0);
DECLARE_ALIGNED(16, int16_t,
im_block[(MAX_SB_SIZE + HORIZ_EXTRA_ROWS) * MAX_SB_SIZE]);
@@ -1092,8 +1120,8 @@
const int16x8_t x_filter = vshrq_n_s16(vld1q_s16(x_filter_ptr), 1);
const int16x8_t y_filter = vld1q_s16(y_filter_ptr);
- dist_wtd_convolve_2d_horiz_neon(src_ptr, src_stride, im_block, im_stride,
- x_filter, im_h, w);
+ dist_wtd_convolve_2d_horiz_8tap_neon(src_ptr, src_stride, im_block, im_stride,
+ x_filter, im_h, w);
if (clamped_y_taps == 6) {
dist_wtd_convolve_2d_vert_6tap_neon(im_block + im_stride, im_stride, dst8,
