limit_v = _mm_unpacklo_epi8(_mm_load_si128((const __m128i *)limit), zero);
thresh_v = _mm_unpacklo_epi8(_mm_load_si128((const __m128i *)thresh), zero);
t80 = _mm_set1_epi16(0x80);
- tff80 = _mm_set1_epi16(0xff80);
- tffe0 = _mm_set1_epi16(0xffe0);
+ tff80 = _mm_set1_epi16((int16_t)0xff80);
+ tffe0 = _mm_set1_epi16((int16_t)0xffe0);
t1f = _mm_srli_epi16(_mm_set1_epi16(0x1fff), 8);
t7f = _mm_srli_epi16(_mm_set1_epi16(0x7fff), 8);
} else if (bd == 10) {
thresh_v = _mm_slli_epi16(
_mm_unpacklo_epi8(_mm_load_si128((const __m128i *)thresh), zero), 2);
t80 = _mm_slli_epi16(_mm_set1_epi16(0x80), 2);
- tff80 = _mm_slli_epi16(_mm_set1_epi16(0xff80), 2);
- tffe0 = _mm_slli_epi16(_mm_set1_epi16(0xffe0), 2);
+ tff80 = _mm_slli_epi16(_mm_set1_epi16((int16_t)0xff80), 2);
+ tffe0 = _mm_slli_epi16(_mm_set1_epi16((int16_t)0xffe0), 2);
t1f = _mm_srli_epi16(_mm_set1_epi16(0x1fff), 6);
t7f = _mm_srli_epi16(_mm_set1_epi16(0x7fff), 6);
} else { // bd == 12
thresh_v = _mm_slli_epi16(
_mm_unpacklo_epi8(_mm_load_si128((const __m128i *)thresh), zero), 4);
t80 = _mm_slli_epi16(_mm_set1_epi16(0x80), 4);
- tff80 = _mm_slli_epi16(_mm_set1_epi16(0xff80), 4);
- tffe0 = _mm_slli_epi16(_mm_set1_epi16(0xffe0), 4);
+ tff80 = _mm_slli_epi16(_mm_set1_epi16((int16_t)0xff80), 4);
+ tffe0 = _mm_slli_epi16(_mm_set1_epi16((int16_t)0xffe0), 4);
t1f = _mm_srli_epi16(_mm_set1_epi16(0x1fff), 4);
t7f = _mm_srli_epi16(_mm_set1_epi16(0x7fff), 4);
}
abs_p1p0 =
_mm_or_si128(_mm_subs_epu8(q1p1, q0p0), _mm_subs_epu8(q0p0, q1p1));
abs_q1q0 = _mm_srli_si128(abs_p1p0, 8);
- fe = _mm_set1_epi8(0xfe);
+ fe = _mm_set1_epi8((int8_t)0xfe);
ff = _mm_cmpeq_epi8(abs_p1p0, abs_p1p0);
abs_p0q0 =
_mm_or_si128(_mm_subs_epu8(q0p0, p0q0), _mm_subs_epu8(p0q0, q0p0));
{
const __m128i t4 = _mm_set1_epi8(4);
const __m128i t3 = _mm_set1_epi8(3);
- const __m128i t80 = _mm_set1_epi8(0x80);
+ const __m128i t80 = _mm_set1_epi8((int8_t)0x80);
const __m128i t1 = _mm_set1_epi16(0x1);
__m128i qs1ps1 = _mm_xor_si128(q1p1, t80);
__m128i qs0ps0 = _mm_xor_si128(q0p0, t80);
_mm_or_si128(_mm_subs_epu8(p1, p0), _mm_subs_epu8(p0, p1));
const __m128i abs_q1q0 =
_mm_or_si128(_mm_subs_epu8(q1, q0), _mm_subs_epu8(q0, q1));
- const __m128i fe = _mm_set1_epi8(0xfe);
+ const __m128i fe = _mm_set1_epi8((int8_t)0xfe);
const __m128i ff = _mm_cmpeq_epi8(abs_p1p0, abs_p1p0);
__m128i abs_p0q0 =
_mm_or_si128(_mm_subs_epu8(p0, q0), _mm_subs_epu8(q0, p0));
{
const __m128i t4 = _mm_set1_epi8(4);
const __m128i t3 = _mm_set1_epi8(3);
- const __m128i t80 = _mm_set1_epi8(0x80);
- const __m128i te0 = _mm_set1_epi8(0xe0);
+ const __m128i t80 = _mm_set1_epi8((int8_t)0x80);
+ const __m128i te0 = _mm_set1_epi8((int8_t)0xe0);
const __m128i t1f = _mm_set1_epi8(0x1f);
const __m128i t1 = _mm_set1_epi8(0x1);
const __m128i t7f = _mm_set1_epi8(0x7f);
do { \
const __m128i t3t4 = \
_mm_set_epi8(3, 3, 3, 3, 3, 3, 3, 3, 4, 4, 4, 4, 4, 4, 4, 4); \
- const __m128i t80 = _mm_set1_epi8(0x80); \
+ const __m128i t80 = _mm_set1_epi8((int8_t)0x80); \
__m128i filter, filter2filter1, work; \
\
ps1ps0 = _mm_xor_si128(p1p0, t80); /* ^ 0x80 */ \
__m128i abs_p1q1, abs_p0q0, abs_q1q0, fe, ff, work;
abs_p1p0 = abs_diff(q1p1, q0p0);
abs_q1q0 = _mm_srli_si128(abs_p1p0, 8);
- fe = _mm_set1_epi8(0xfe);
+ fe = _mm_set1_epi8((int8_t)0xfe);
ff = _mm_cmpeq_epi8(abs_p1p0, abs_p1p0);
abs_p0q0 = abs_diff(q0p0, p0q0);
abs_p1q1 = abs_diff(q1p1, p1q1);
{
const __m128i t4 = _mm_set1_epi8(4);
const __m128i t3 = _mm_set1_epi8(3);
- const __m128i t80 = _mm_set1_epi8(0x80);
+ const __m128i t80 = _mm_set1_epi8((int8_t)0x80);
const __m128i t1 = _mm_set1_epi16(0x1);
__m128i qs1ps1 = _mm_xor_si128(q1p1, t80);
__m128i qs0ps0 = _mm_xor_si128(q0p0, t80);
{
const __m128i abs_p1p0 = abs_diff(p1, p0);
const __m128i abs_q1q0 = abs_diff(q1, q0);
- const __m128i fe = _mm_set1_epi8(0xfe);
+ const __m128i fe = _mm_set1_epi8((int8_t)0xfe);
const __m128i ff = _mm_cmpeq_epi8(zero, zero);
__m128i abs_p0q0 = abs_diff(p0, q0);
__m128i abs_p1q1 = abs_diff(p1, q1);
{
const __m128i t4 = _mm_set1_epi8(4);
const __m128i t3 = _mm_set1_epi8(3);
- const __m128i t80 = _mm_set1_epi8(0x80);
- const __m128i te0 = _mm_set1_epi8(0xe0);
+ const __m128i t80 = _mm_set1_epi8((int8_t)0x80);
+ const __m128i te0 = _mm_set1_epi8((int8_t)0xe0);
const __m128i t1f = _mm_set1_epi8(0x1f);
const __m128i t1 = _mm_set1_epi8(0x1);
const __m128i t7f = _mm_set1_epi8(0x7f);
{
// filter_mask and hev_mask
const __m128i one = _mm_set1_epi8(1);
- const __m128i fe = _mm_set1_epi8(0xfe);
+ const __m128i fe = _mm_set1_epi8((int8_t)0xfe);
const __m128i ff = _mm_cmpeq_epi8(fe, fe);
__m128i abs_p1q1, abs_p0q0, abs_q1q0, abs_p1p0, work;
abs_p1p0 = abs_diff(q1p1, q0p0);
{
const __m128i t4 = _mm_set1_epi8(4);
const __m128i t3 = _mm_set1_epi8(3);
- const __m128i t80 = _mm_set1_epi8(0x80);
+ const __m128i t80 = _mm_set1_epi8((int8_t)0x80);
const __m128i t1 = _mm_set1_epi8(0x1);
const __m128i ps1 =
_mm_xor_si128(_mm_loadl_epi64((__m128i *)(s - 2 * pitch)), t80);
const __m128i abs_q1q0 =
_mm_or_si128(_mm_subs_epu8(q1, q0), _mm_subs_epu8(q0, q1));
const __m128i one = _mm_set1_epi8(1);
- const __m128i fe = _mm_set1_epi8(0xfe);
+ const __m128i fe = _mm_set1_epi8((int8_t)0xfe);
const __m128i ff = _mm_cmpeq_epi8(abs_p1p0, abs_p1p0);
__m128i abs_p0q0 =
_mm_or_si128(_mm_subs_epu8(p0, q0), _mm_subs_epu8(q0, p0));
{
const __m128i t4 = _mm_set1_epi8(4);
const __m128i t3 = _mm_set1_epi8(3);
- const __m128i t80 = _mm_set1_epi8(0x80);
- const __m128i te0 = _mm_set1_epi8(0xe0);
+ const __m128i t80 = _mm_set1_epi8((int8_t)0x80);
+ const __m128i te0 = _mm_set1_epi8((int8_t)0xe0);
const __m128i t1f = _mm_set1_epi8(0x1f);
const __m128i t1 = _mm_set1_epi8(0x1);
const __m128i t7f = _mm_set1_epi8(0x7f);
_mm_or_si128(_mm_subs_epu8(p1, p0), _mm_subs_epu8(p0, p1));
const __m128i abs_q1q0 =
_mm_or_si128(_mm_subs_epu8(q1, q0), _mm_subs_epu8(q0, q1));
- const __m128i fe = _mm_set1_epi8(0xfe);
+ const __m128i fe = _mm_set1_epi8((int8_t)0xfe);
const __m128i ff = _mm_cmpeq_epi8(abs_p1p0, abs_p1p0);
__m128i abs_p0q0 =
_mm_or_si128(_mm_subs_epu8(p0, q0), _mm_subs_epu8(q0, p0));
{
const __m128i t4 = _mm_set1_epi8(4);
const __m128i t3 = _mm_set1_epi8(3);
- const __m128i t80 = _mm_set1_epi8(0x80);
- const __m128i te0 = _mm_set1_epi8(0xe0);
+ const __m128i t80 = _mm_set1_epi8((int8_t)0x80);
+ const __m128i te0 = _mm_set1_epi8((int8_t)0xe0);
const __m128i t1f = _mm_set1_epi8(0x1f);
const __m128i t1 = _mm_set1_epi8(0x1);
const __m128i t7f = _mm_set1_epi8(0x7f);