Merge changes Iebe9842f,I174b67a5,I80ed1a16

* changes: vpx_convolve_neon: prefer != 0 to > 0 in tests vpx_convolve_avg_neon: prefer != 0 to > 0 in tests vpx_convolve_copy_neon: prefer != 0 to > 0 in tests

James Zern committed 5 years ago

Commit c4fceb92 parents 5564337 ff67c843 files · +18 −18Browse files

Unified · Split

3 files

vpx_dsp/arm/vpx_convolve8_neon.c +10 −10
vpx_dsp/arm/vpx_convolve_avg_neon.c +4 −4
vpx_dsp/arm/vpx_convolve_copy_neon.c +4 −4

vpx_dsp/arm/vpx_convolve8_neon.c+10−10

@@ -145,7 +145,7 @@ void vpx_convolve8_horiz_neon(const uint8_t *src, ptrdiff_t src_stride, · 144 unmodified lines
145	145	src += 4;
146	146	dst += 4;
147	147	w -= 4;
148		− } while (w > 0);
	148	+ } while (w != 0);
149	149	} else {
150	150	const int16x8_t filter3 = vdupq_lane_s16(vget_low_s16(filters), 3);
151	151	const int16x8_t filter4 = vdupq_lane_s16(vget_high_s16(filters), 0);
@@ -296,7 +296,7 @@ void vpx_convolve8_horiz_neon(const uint8_t *src, ptrdiff_t src_stride, · 144 unmodified lines
296	296	s += 8;
297	297	d += 8;
298	298	width -= 8;
299		− } while (width > 0);
	299	+ } while (width != 0);
300	300	src += 8 * src_stride;
301	301	dst += 8 * dst_stride;
302	302	h -= 8;
@@ -402,7 +402,7 @@ void vpx_convolve8_avg_horiz_neon(const uint8_t *src, ptrdiff_t src_stride, · 99 unmodified lines
402	402	src += 4;
403	403	dst += 4;
404	404	w -= 4;
405		− } while (w > 0);
	405	+ } while (w != 0);
406	406	} else {
407	407	const int16x8_t filter3 = vdupq_lane_s16(vget_low_s16(filters), 3);
408	408	const int16x8_t filter4 = vdupq_lane_s16(vget_high_s16(filters), 0);
@@ -586,7 +586,7 @@ void vpx_convolve8_avg_horiz_neon(const uint8_t *src, ptrdiff_t src_stride, · 177 unmodified lines
586	586	s += 8;
587	587	d += 8;
588	588	width -= 8;
589		− } while (width > 0);
	589	+ } while (width != 0);
590	590	src += 8 * src_stride;
591	591	dst += 8 * dst_stride;
592	592	h -= 8;
@@ -679,7 +679,7 @@ void vpx_convolve8_vert_neon(const uint8_t *src, ptrdiff_t src_stride, · 86 unmodified lines
679	679	s5 = s9;
680	680	s6 = s10;
681	681	h -= 4;
682		− } while (h > 0);
	682	+ } while (h != 0);
683	683	} else {
684	684	const int16x8_t filter3 = vdupq_lane_s16(vget_low_s16(filters), 3);
685	685	const int16x8_t filter4 = vdupq_lane_s16(vget_high_s16(filters), 0);
@@ -759,11 +759,11 @@ void vpx_convolve8_vert_neon(const uint8_t *src, ptrdiff_t src_stride, · 73 unmodified lines
759	759	s5 = s9;
760	760	s6 = s10;
761	761	height -= 4;
762		− } while (height > 0);
	762	+ } while (height != 0);
763	763	src += 8;
764	764	dst += 8;
765	765	w -= 8;
766		− } while (w > 0);
	766	+ } while (w != 0);
767	767	}
768	768	}
769	769
@@ -860,7 +860,7 @@ void vpx_convolve8_avg_vert_neon(const uint8_t *src, ptrdiff_t src_stride, · 90 unmodified lines
860	860	s5 = s9;
861	861	s6 = s10;
862	862	h -= 4;
863		− } while (h > 0);
	863	+ } while (h != 0);
864	864	} else {
865	865	const int16x8_t filter3 = vdupq_lane_s16(vget_low_s16(filters), 3);
866	866	const int16x8_t filter4 = vdupq_lane_s16(vget_high_s16(filters), 0);
@@ -950,10 +950,10 @@ void vpx_convolve8_avg_vert_neon(const uint8_t *src, ptrdiff_t src_stride, · 83 unmodified lines
950	950	s5 = s9;
951	951	s6 = s10;
952	952	height -= 4;
953		− } while (height > 0);
	953	+ } while (height != 0);
954	954	src += 8;
955	955	dst += 8;
956	956	w -= 8;
957		− } while (w > 0);
	957	+ } while (w != 0);
958	958	}
959	959	}

vpx_dsp/arm/vpx_convolve_avg_neon.c+4−4

@@ -43,7 +43,7 @@ void vpx_convolve_avg_neon(const uint8_t *src, ptrdiff_t src_stride, · 42 unmodified lines
43	43	vst1_lane_u32((uint32_t *)dst, vreinterpret_u32_u8(dd0), 1);
44	44	dst += dst_stride;
45	45	h -= 2;
46		− } while (h > 0);
	46	+ } while (h != 0);
47	47	} else if (w == 8) { // avg8
48	48	uint8x8_t s0, s1, d0, d1;
49	49	uint8x16_t s01, d01;
@@ -64,7 +64,7 @@ void vpx_convolve_avg_neon(const uint8_t *src, ptrdiff_t src_stride, · 14 unmodified lines
64	64	vst1_u8(dst, vget_high_u8(d01));
65	65	dst += dst_stride;
66	66	h -= 2;
67		− } while (h > 0);
	67	+ } while (h != 0);
68	68	} else if (w < 32) { // avg16
69	69	uint8x16_t s0, s1, d0, d1;
70	70	do {
@@ -83,7 +83,7 @@ void vpx_convolve_avg_neon(const uint8_t *src, ptrdiff_t src_stride, · 12 unmodified lines
83	83	vst1q_u8(dst, d1);
84	84	dst += dst_stride;
85	85	h -= 2;
86		− } while (h > 0);
	86	+ } while (h != 0);
87	87	} else if (w == 32) { // avg32
88	88	uint8x16_t s0, s1, s2, s3, d0, d1, d2, d3;
89	89	do {
@@ -110,7 +110,7 @@ void vpx_convolve_avg_neon(const uint8_t *src, ptrdiff_t src_stride, · 20 unmodified lines
110	110	vst1q_u8(dst + 16, d3);
111	111	dst += dst_stride;
112	112	h -= 2;
113		− } while (h > 0);
	113	+ } while (h != 0);
114	114	} else { // avg64
115	115	uint8x16_t s0, s1, s2, s3, d0, d1, d2, d3;
116	116	do {

vpx_dsp/arm/vpx_convolve_copy_neon.c+4−4

@@ -33,7 +33,7 @@ void vpx_convolve_copy_neon(const uint8_t *src, ptrdiff_t src_stride, · 32 unmodified lines
33	33	src += src_stride;
34	34	dst += dst_stride;
35	35	h -= 2;
36		− } while (h > 0);
	36	+ } while (h != 0);
37	37	} else if (w == 8) { // copy8
38	38	uint8x8_t s0, s1;
39	39	do {
@@ -47,7 +47,7 @@ void vpx_convolve_copy_neon(const uint8_t *src, ptrdiff_t src_stride, · 7 unmodified lines
47	47	vst1_u8(dst, s1);
48	48	dst += dst_stride;
49	49	h -= 2;
50		− } while (h > 0);
	50	+ } while (h != 0);
51	51	} else if (w < 32) { // copy16
52	52	uint8x16_t s0, s1;
53	53	do {
@@ -61,7 +61,7 @@ void vpx_convolve_copy_neon(const uint8_t *src, ptrdiff_t src_stride, · 7 unmodified lines
61	61	vst1q_u8(dst, s1);
62	62	dst += dst_stride;
63	63	h -= 2;
64		− } while (h > 0);
	64	+ } while (h != 0);
65	65	} else if (w == 32) { // copy32
66	66	uint8x16_t s0, s1, s2, s3;
67	67	do {
@@ -79,7 +79,7 @@ void vpx_convolve_copy_neon(const uint8_t *src, ptrdiff_t src_stride, · 11 unmodified lines
79	79	vst1q_u8(dst + 16, s3);
80	80	dst += dst_stride;
81	81	h -= 2;
82		− } while (h > 0);
	82	+ } while (h != 0);
83	83	} else { // copy64
84	84	uint8x16_t s0, s1, s2, s3;
85	85	do {

Merge changes Iebe9842f,I174b67a5,I80ed1a16

Keyboard shortcuts

General

Issues and pull requests

Issues, pull requests and Inbox

Repository code