Home
last modified time | relevance | path

Searched refs:_mm_sub_epi64 (Results 1 – 25 of 90) sorted by relevance

1234

/external/libvpx/libvpx/vp9/common/x86/
Dvp9_highbd_iht16x16_add_sse4.c52 s1[0] = _mm_sub_epi64(t10[0], t01[0]); in highbd_iadst_butterfly_sse4_1()
53 s1[1] = _mm_sub_epi64(t10[1], t01[1]); in highbd_iadst_butterfly_sse4_1()
91 x8[0] = _mm_sub_epi64(s0[0], s8[0]); in highbd_iadst16_4col_sse4_1()
92 x8[1] = _mm_sub_epi64(s0[1], s8[1]); in highbd_iadst16_4col_sse4_1()
93 x9[0] = _mm_sub_epi64(s1[0], s9[0]); in highbd_iadst16_4col_sse4_1()
94 x9[1] = _mm_sub_epi64(s1[1], s9[1]); in highbd_iadst16_4col_sse4_1()
95 x10[0] = _mm_sub_epi64(s2[0], s10[0]); in highbd_iadst16_4col_sse4_1()
96 x10[1] = _mm_sub_epi64(s2[1], s10[1]); in highbd_iadst16_4col_sse4_1()
97 x11[0] = _mm_sub_epi64(s3[0], s11[0]); in highbd_iadst16_4col_sse4_1()
98 x11[1] = _mm_sub_epi64(s3[1], s11[1]); in highbd_iadst16_4col_sse4_1()
[all …]
Dvp9_highbd_iht8x8_add_sse4.c52 s1[0] = _mm_sub_epi64(t10[0], t01[0]); in highbd_iadst_butterfly_sse4_1()
53 s1[1] = _mm_sub_epi64(t10[1], t01[1]); in highbd_iadst_butterfly_sse4_1()
69 x4[0] = _mm_sub_epi64(s0[0], s4[0]); in highbd_iadst8_sse4_1()
70 x4[1] = _mm_sub_epi64(s0[1], s4[1]); in highbd_iadst8_sse4_1()
71 x5[0] = _mm_sub_epi64(s1[0], s5[0]); in highbd_iadst8_sse4_1()
72 x5[1] = _mm_sub_epi64(s1[1], s5[1]); in highbd_iadst8_sse4_1()
80 x6[0] = _mm_sub_epi64(s2[0], s6[0]); in highbd_iadst8_sse4_1()
81 x6[1] = _mm_sub_epi64(s2[1], s6[1]); in highbd_iadst8_sse4_1()
82 x7[0] = _mm_sub_epi64(s3[0], s7[0]); in highbd_iadst8_sse4_1()
83 x7[1] = _mm_sub_epi64(s3[1], s7[1]); in highbd_iadst8_sse4_1()
[all …]
Dvp9_highbd_iht4x4_add_sse4.c54 t1[0] = _mm_sub_epi64(s1[0], s4[0]); in highbd_iadst4_sse4_1()
55 t1[1] = _mm_sub_epi64(s1[1], s4[1]); in highbd_iadst4_sse4_1()
56 t1[0] = _mm_sub_epi64(t1[0], s6[0]); in highbd_iadst4_sse4_1()
57 t1[1] = _mm_sub_epi64(t1[1], s6[1]); in highbd_iadst4_sse4_1()
70 s3[0] = _mm_sub_epi64(s3[0], s2[0]); in highbd_iadst4_sse4_1()
71 s3[1] = _mm_sub_epi64(s3[1], s2[1]); in highbd_iadst4_sse4_1()
/external/XNNPACK/src/qu8-requantization/
Dq31-ssse3.c87 …const __m128i x_product_even = _mm_sub_epi64(_mm_xor_si128(x_abs_product_even, x_neg_mask_even), x… in xnn_qu8_requantize_q31__ssse3()
88 …const __m128i y_product_even = _mm_sub_epi64(_mm_xor_si128(y_abs_product_even, y_neg_mask_even), y… in xnn_qu8_requantize_q31__ssse3()
89 …const __m128i z_product_even = _mm_sub_epi64(_mm_xor_si128(z_abs_product_even, z_neg_mask_even), z… in xnn_qu8_requantize_q31__ssse3()
90 …const __m128i w_product_even = _mm_sub_epi64(_mm_xor_si128(w_abs_product_even, w_neg_mask_even), w… in xnn_qu8_requantize_q31__ssse3()
107 …const __m128i x_product_odd = _mm_sub_epi64(_mm_xor_si128(x_abs_product_odd, x_neg_mask_odd), x_ne… in xnn_qu8_requantize_q31__ssse3()
108 …const __m128i y_product_odd = _mm_sub_epi64(_mm_xor_si128(y_abs_product_odd, y_neg_mask_odd), y_ne… in xnn_qu8_requantize_q31__ssse3()
109 …const __m128i z_product_odd = _mm_sub_epi64(_mm_xor_si128(z_abs_product_odd, z_neg_mask_odd), z_ne… in xnn_qu8_requantize_q31__ssse3()
110 …const __m128i w_product_odd = _mm_sub_epi64(_mm_xor_si128(w_abs_product_odd, w_neg_mask_odd), w_ne… in xnn_qu8_requantize_q31__ssse3()
Dq31-sse2.c87 …const __m128i x_product_even = _mm_sub_epi64(_mm_xor_si128(x_abs_product_even, x_neg_mask_even), x… in xnn_qu8_requantize_q31__sse2()
88 …const __m128i y_product_even = _mm_sub_epi64(_mm_xor_si128(y_abs_product_even, y_neg_mask_even), y… in xnn_qu8_requantize_q31__sse2()
89 …const __m128i z_product_even = _mm_sub_epi64(_mm_xor_si128(z_abs_product_even, z_neg_mask_even), z… in xnn_qu8_requantize_q31__sse2()
90 …const __m128i w_product_even = _mm_sub_epi64(_mm_xor_si128(w_abs_product_even, w_neg_mask_even), w… in xnn_qu8_requantize_q31__sse2()
107 …const __m128i x_product_odd = _mm_sub_epi64(_mm_xor_si128(x_abs_product_odd, x_neg_mask_odd), x_ne… in xnn_qu8_requantize_q31__sse2()
108 …const __m128i y_product_odd = _mm_sub_epi64(_mm_xor_si128(y_abs_product_odd, y_neg_mask_odd), y_ne… in xnn_qu8_requantize_q31__sse2()
109 …const __m128i z_product_odd = _mm_sub_epi64(_mm_xor_si128(z_abs_product_odd, z_neg_mask_odd), z_ne… in xnn_qu8_requantize_q31__sse2()
110 …const __m128i w_product_odd = _mm_sub_epi64(_mm_xor_si128(w_abs_product_odd, w_neg_mask_odd), w_ne… in xnn_qu8_requantize_q31__sse2()
/external/XNNPACK/src/qs8-requantization/
Dq31-ssse3.c87 …const __m128i x_product_even = _mm_sub_epi64(_mm_xor_si128(x_abs_product_even, x_neg_mask_even), x… in xnn_qs8_requantize_q31__ssse3()
88 …const __m128i y_product_even = _mm_sub_epi64(_mm_xor_si128(y_abs_product_even, y_neg_mask_even), y… in xnn_qs8_requantize_q31__ssse3()
89 …const __m128i z_product_even = _mm_sub_epi64(_mm_xor_si128(z_abs_product_even, z_neg_mask_even), z… in xnn_qs8_requantize_q31__ssse3()
90 …const __m128i w_product_even = _mm_sub_epi64(_mm_xor_si128(w_abs_product_even, w_neg_mask_even), w… in xnn_qs8_requantize_q31__ssse3()
107 …const __m128i x_product_odd = _mm_sub_epi64(_mm_xor_si128(x_abs_product_odd, x_neg_mask_odd), x_ne… in xnn_qs8_requantize_q31__ssse3()
108 …const __m128i y_product_odd = _mm_sub_epi64(_mm_xor_si128(y_abs_product_odd, y_neg_mask_odd), y_ne… in xnn_qs8_requantize_q31__ssse3()
109 …const __m128i z_product_odd = _mm_sub_epi64(_mm_xor_si128(z_abs_product_odd, z_neg_mask_odd), z_ne… in xnn_qs8_requantize_q31__ssse3()
110 …const __m128i w_product_odd = _mm_sub_epi64(_mm_xor_si128(w_abs_product_odd, w_neg_mask_odd), w_ne… in xnn_qs8_requantize_q31__ssse3()
Dq31-sse2.c87 …const __m128i x_product_even = _mm_sub_epi64(_mm_xor_si128(x_abs_product_even, x_neg_mask_even), x… in xnn_qs8_requantize_q31__sse2()
88 …const __m128i y_product_even = _mm_sub_epi64(_mm_xor_si128(y_abs_product_even, y_neg_mask_even), y… in xnn_qs8_requantize_q31__sse2()
89 …const __m128i z_product_even = _mm_sub_epi64(_mm_xor_si128(z_abs_product_even, z_neg_mask_even), z… in xnn_qs8_requantize_q31__sse2()
90 …const __m128i w_product_even = _mm_sub_epi64(_mm_xor_si128(w_abs_product_even, w_neg_mask_even), w… in xnn_qs8_requantize_q31__sse2()
107 …const __m128i x_product_odd = _mm_sub_epi64(_mm_xor_si128(x_abs_product_odd, x_neg_mask_odd), x_ne… in xnn_qs8_requantize_q31__sse2()
108 …const __m128i y_product_odd = _mm_sub_epi64(_mm_xor_si128(y_abs_product_odd, y_neg_mask_odd), y_ne… in xnn_qs8_requantize_q31__sse2()
109 …const __m128i z_product_odd = _mm_sub_epi64(_mm_xor_si128(z_abs_product_odd, z_neg_mask_odd), z_ne… in xnn_qs8_requantize_q31__sse2()
110 …const __m128i w_product_odd = _mm_sub_epi64(_mm_xor_si128(w_abs_product_odd, w_neg_mask_odd), w_ne… in xnn_qs8_requantize_q31__sse2()
/external/XNNPACK/src/qs8-gemm/gen/
D3x4c8-xw-minmax-ssse3.c141 const __m128i vprod0x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x02, vnmask0x02), vnmask0x02); in xnn_qs8_gemm_xw_minmax_ukernel_3x4c8__ssse3()
142 const __m128i vprod1x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x02, vnmask1x02), vnmask1x02); in xnn_qs8_gemm_xw_minmax_ukernel_3x4c8__ssse3()
143 const __m128i vprod2x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x02, vnmask2x02), vnmask2x02); in xnn_qs8_gemm_xw_minmax_ukernel_3x4c8__ssse3()
157 const __m128i vprod0x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x13, vnmask0x13), vnmask0x13); in xnn_qs8_gemm_xw_minmax_ukernel_3x4c8__ssse3()
158 const __m128i vprod1x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x13, vnmask1x13), vnmask1x13); in xnn_qs8_gemm_xw_minmax_ukernel_3x4c8__ssse3()
159 const __m128i vprod2x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x13, vnmask2x13), vnmask2x13); in xnn_qs8_gemm_xw_minmax_ukernel_3x4c8__ssse3()
D4x4c2-xw-minmax-sse2.c210 const __m128i vprod0x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x02, vnmask0x02), vnmask0x02); in xnn_qs8_gemm_xw_minmax_ukernel_4x4c2__sse2()
211 const __m128i vprod1x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x02, vnmask1x02), vnmask1x02); in xnn_qs8_gemm_xw_minmax_ukernel_4x4c2__sse2()
212 const __m128i vprod2x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x02, vnmask2x02), vnmask2x02); in xnn_qs8_gemm_xw_minmax_ukernel_4x4c2__sse2()
213 const __m128i vprod3x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod3x02, vnmask3x02), vnmask3x02); in xnn_qs8_gemm_xw_minmax_ukernel_4x4c2__sse2()
230 const __m128i vprod0x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x13, vnmask0x13), vnmask0x13); in xnn_qs8_gemm_xw_minmax_ukernel_4x4c2__sse2()
231 const __m128i vprod1x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x13, vnmask1x13), vnmask1x13); in xnn_qs8_gemm_xw_minmax_ukernel_4x4c2__sse2()
232 const __m128i vprod2x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x13, vnmask2x13), vnmask2x13); in xnn_qs8_gemm_xw_minmax_ukernel_4x4c2__sse2()
233 const __m128i vprod3x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod3x13, vnmask3x13), vnmask3x13); in xnn_qs8_gemm_xw_minmax_ukernel_4x4c2__sse2()
D4x4c2-minmax-sse2-ld128.c217 const __m128i vprod0x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x02, vnmask0x02), vnmask0x02); in xnn_qs8_gemm_minmax_ukernel_4x4c2__sse2_ld128()
218 const __m128i vprod1x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x02, vnmask1x02), vnmask1x02); in xnn_qs8_gemm_minmax_ukernel_4x4c2__sse2_ld128()
219 const __m128i vprod2x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x02, vnmask2x02), vnmask2x02); in xnn_qs8_gemm_minmax_ukernel_4x4c2__sse2_ld128()
220 const __m128i vprod3x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod3x02, vnmask3x02), vnmask3x02); in xnn_qs8_gemm_minmax_ukernel_4x4c2__sse2_ld128()
237 const __m128i vprod0x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x13, vnmask0x13), vnmask0x13); in xnn_qs8_gemm_minmax_ukernel_4x4c2__sse2_ld128()
238 const __m128i vprod1x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x13, vnmask1x13), vnmask1x13); in xnn_qs8_gemm_minmax_ukernel_4x4c2__sse2_ld128()
239 const __m128i vprod2x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x13, vnmask2x13), vnmask2x13); in xnn_qs8_gemm_minmax_ukernel_4x4c2__sse2_ld128()
240 const __m128i vprod3x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod3x13, vnmask3x13), vnmask3x13); in xnn_qs8_gemm_minmax_ukernel_4x4c2__sse2_ld128()
D4x4c2-minmax-sse2-ld64.c217 const __m128i vprod0x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x02, vnmask0x02), vnmask0x02); in xnn_qs8_gemm_minmax_ukernel_4x4c2__sse2_ld64()
218 const __m128i vprod1x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x02, vnmask1x02), vnmask1x02); in xnn_qs8_gemm_minmax_ukernel_4x4c2__sse2_ld64()
219 const __m128i vprod2x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x02, vnmask2x02), vnmask2x02); in xnn_qs8_gemm_minmax_ukernel_4x4c2__sse2_ld64()
220 const __m128i vprod3x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod3x02, vnmask3x02), vnmask3x02); in xnn_qs8_gemm_minmax_ukernel_4x4c2__sse2_ld64()
237 const __m128i vprod0x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x13, vnmask0x13), vnmask0x13); in xnn_qs8_gemm_minmax_ukernel_4x4c2__sse2_ld64()
238 const __m128i vprod1x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x13, vnmask1x13), vnmask1x13); in xnn_qs8_gemm_minmax_ukernel_4x4c2__sse2_ld64()
239 const __m128i vprod2x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x13, vnmask2x13), vnmask2x13); in xnn_qs8_gemm_minmax_ukernel_4x4c2__sse2_ld64()
240 const __m128i vprod3x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod3x13, vnmask3x13), vnmask3x13); in xnn_qs8_gemm_minmax_ukernel_4x4c2__sse2_ld64()
D4x4c2-minmax-ssse3-ld64.c217 const __m128i vprod0x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x02, vnmask0x02), vnmask0x02); in xnn_qs8_gemm_minmax_ukernel_4x4c2__ssse3_ld64()
218 const __m128i vprod1x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x02, vnmask1x02), vnmask1x02); in xnn_qs8_gemm_minmax_ukernel_4x4c2__ssse3_ld64()
219 const __m128i vprod2x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x02, vnmask2x02), vnmask2x02); in xnn_qs8_gemm_minmax_ukernel_4x4c2__ssse3_ld64()
220 const __m128i vprod3x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod3x02, vnmask3x02), vnmask3x02); in xnn_qs8_gemm_minmax_ukernel_4x4c2__ssse3_ld64()
237 const __m128i vprod0x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x13, vnmask0x13), vnmask0x13); in xnn_qs8_gemm_minmax_ukernel_4x4c2__ssse3_ld64()
238 const __m128i vprod1x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x13, vnmask1x13), vnmask1x13); in xnn_qs8_gemm_minmax_ukernel_4x4c2__ssse3_ld64()
239 const __m128i vprod2x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x13, vnmask2x13), vnmask2x13); in xnn_qs8_gemm_minmax_ukernel_4x4c2__ssse3_ld64()
240 const __m128i vprod3x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod3x13, vnmask3x13), vnmask3x13); in xnn_qs8_gemm_minmax_ukernel_4x4c2__ssse3_ld64()
D4x4c2-xw-minmax-ssse3.c210 const __m128i vprod0x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x02, vnmask0x02), vnmask0x02); in xnn_qs8_gemm_xw_minmax_ukernel_4x4c2__ssse3()
211 const __m128i vprod1x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x02, vnmask1x02), vnmask1x02); in xnn_qs8_gemm_xw_minmax_ukernel_4x4c2__ssse3()
212 const __m128i vprod2x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x02, vnmask2x02), vnmask2x02); in xnn_qs8_gemm_xw_minmax_ukernel_4x4c2__ssse3()
213 const __m128i vprod3x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod3x02, vnmask3x02), vnmask3x02); in xnn_qs8_gemm_xw_minmax_ukernel_4x4c2__ssse3()
230 const __m128i vprod0x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x13, vnmask0x13), vnmask0x13); in xnn_qs8_gemm_xw_minmax_ukernel_4x4c2__ssse3()
231 const __m128i vprod1x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x13, vnmask1x13), vnmask1x13); in xnn_qs8_gemm_xw_minmax_ukernel_4x4c2__ssse3()
232 const __m128i vprod2x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x13, vnmask2x13), vnmask2x13); in xnn_qs8_gemm_xw_minmax_ukernel_4x4c2__ssse3()
233 const __m128i vprod3x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod3x13, vnmask3x13), vnmask3x13); in xnn_qs8_gemm_xw_minmax_ukernel_4x4c2__ssse3()
D4x4c2-minmax-ssse3-ld128.c217 const __m128i vprod0x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x02, vnmask0x02), vnmask0x02); in xnn_qs8_gemm_minmax_ukernel_4x4c2__ssse3_ld128()
218 const __m128i vprod1x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x02, vnmask1x02), vnmask1x02); in xnn_qs8_gemm_minmax_ukernel_4x4c2__ssse3_ld128()
219 const __m128i vprod2x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x02, vnmask2x02), vnmask2x02); in xnn_qs8_gemm_minmax_ukernel_4x4c2__ssse3_ld128()
220 const __m128i vprod3x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod3x02, vnmask3x02), vnmask3x02); in xnn_qs8_gemm_minmax_ukernel_4x4c2__ssse3_ld128()
237 const __m128i vprod0x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x13, vnmask0x13), vnmask0x13); in xnn_qs8_gemm_minmax_ukernel_4x4c2__ssse3_ld128()
238 const __m128i vprod1x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x13, vnmask1x13), vnmask1x13); in xnn_qs8_gemm_minmax_ukernel_4x4c2__ssse3_ld128()
239 const __m128i vprod2x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x13, vnmask2x13), vnmask2x13); in xnn_qs8_gemm_minmax_ukernel_4x4c2__ssse3_ld128()
240 const __m128i vprod3x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod3x13, vnmask3x13), vnmask3x13); in xnn_qs8_gemm_minmax_ukernel_4x4c2__ssse3_ld128()
D3x4c8-minmax-sse2-ld64.c145 const __m128i vprod0x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x02, vnmask0x02), vnmask0x02); in xnn_qs8_gemm_minmax_ukernel_3x4c8__sse2_ld64()
146 const __m128i vprod1x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x02, vnmask1x02), vnmask1x02); in xnn_qs8_gemm_minmax_ukernel_3x4c8__sse2_ld64()
147 const __m128i vprod2x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x02, vnmask2x02), vnmask2x02); in xnn_qs8_gemm_minmax_ukernel_3x4c8__sse2_ld64()
161 const __m128i vprod0x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x13, vnmask0x13), vnmask0x13); in xnn_qs8_gemm_minmax_ukernel_3x4c8__sse2_ld64()
162 const __m128i vprod1x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x13, vnmask1x13), vnmask1x13); in xnn_qs8_gemm_minmax_ukernel_3x4c8__sse2_ld64()
163 const __m128i vprod2x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x13, vnmask2x13), vnmask2x13); in xnn_qs8_gemm_minmax_ukernel_3x4c8__sse2_ld64()
D3x4c8-minmax-sse2-ld128.c143 const __m128i vprod0x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x02, vnmask0x02), vnmask0x02); in xnn_qs8_gemm_minmax_ukernel_3x4c8__sse2_ld128()
144 const __m128i vprod1x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x02, vnmask1x02), vnmask1x02); in xnn_qs8_gemm_minmax_ukernel_3x4c8__sse2_ld128()
145 const __m128i vprod2x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x02, vnmask2x02), vnmask2x02); in xnn_qs8_gemm_minmax_ukernel_3x4c8__sse2_ld128()
159 const __m128i vprod0x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x13, vnmask0x13), vnmask0x13); in xnn_qs8_gemm_minmax_ukernel_3x4c8__sse2_ld128()
160 const __m128i vprod1x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x13, vnmask1x13), vnmask1x13); in xnn_qs8_gemm_minmax_ukernel_3x4c8__sse2_ld128()
161 const __m128i vprod2x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x13, vnmask2x13), vnmask2x13); in xnn_qs8_gemm_minmax_ukernel_3x4c8__sse2_ld128()
D3x4c8-minmax-ssse3-ld128.c143 const __m128i vprod0x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x02, vnmask0x02), vnmask0x02); in xnn_qs8_gemm_minmax_ukernel_3x4c8__ssse3_ld128()
144 const __m128i vprod1x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x02, vnmask1x02), vnmask1x02); in xnn_qs8_gemm_minmax_ukernel_3x4c8__ssse3_ld128()
145 const __m128i vprod2x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x02, vnmask2x02), vnmask2x02); in xnn_qs8_gemm_minmax_ukernel_3x4c8__ssse3_ld128()
159 const __m128i vprod0x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x13, vnmask0x13), vnmask0x13); in xnn_qs8_gemm_minmax_ukernel_3x4c8__ssse3_ld128()
160 const __m128i vprod1x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x13, vnmask1x13), vnmask1x13); in xnn_qs8_gemm_minmax_ukernel_3x4c8__ssse3_ld128()
161 const __m128i vprod2x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x13, vnmask2x13), vnmask2x13); in xnn_qs8_gemm_minmax_ukernel_3x4c8__ssse3_ld128()
/external/XNNPACK/src/qu8-gemm/
D4x4c2-minmax-sse2.c222 const __m128i vprod0x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x02, vnmask0x02), vnmask0x02); in xnn_qu8_gemm_minmax_ukernel_4x4c2__sse2()
223 const __m128i vprod1x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x02, vnmask1x02), vnmask1x02); in xnn_qu8_gemm_minmax_ukernel_4x4c2__sse2()
224 const __m128i vprod2x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x02, vnmask2x02), vnmask2x02); in xnn_qu8_gemm_minmax_ukernel_4x4c2__sse2()
225 const __m128i vprod3x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod3x02, vnmask3x02), vnmask3x02); in xnn_qu8_gemm_minmax_ukernel_4x4c2__sse2()
242 const __m128i vprod0x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x13, vnmask0x13), vnmask0x13); in xnn_qu8_gemm_minmax_ukernel_4x4c2__sse2()
243 const __m128i vprod1x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x13, vnmask1x13), vnmask1x13); in xnn_qu8_gemm_minmax_ukernel_4x4c2__sse2()
244 const __m128i vprod2x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x13, vnmask2x13), vnmask2x13); in xnn_qu8_gemm_minmax_ukernel_4x4c2__sse2()
245 const __m128i vprod3x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod3x13, vnmask3x13), vnmask3x13); in xnn_qu8_gemm_minmax_ukernel_4x4c2__sse2()
/external/XNNPACK/src/qs8-igemm/gen/
D4x4c2-minmax-sse2-ld128.c236 const __m128i vprod0x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x02, vnmask0x02), vnmask0x02); in xnn_qs8_igemm_minmax_ukernel_4x4c2__sse2_ld128()
237 const __m128i vprod1x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x02, vnmask1x02), vnmask1x02); in xnn_qs8_igemm_minmax_ukernel_4x4c2__sse2_ld128()
238 const __m128i vprod2x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x02, vnmask2x02), vnmask2x02); in xnn_qs8_igemm_minmax_ukernel_4x4c2__sse2_ld128()
239 const __m128i vprod3x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod3x02, vnmask3x02), vnmask3x02); in xnn_qs8_igemm_minmax_ukernel_4x4c2__sse2_ld128()
256 const __m128i vprod0x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x13, vnmask0x13), vnmask0x13); in xnn_qs8_igemm_minmax_ukernel_4x4c2__sse2_ld128()
257 const __m128i vprod1x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x13, vnmask1x13), vnmask1x13); in xnn_qs8_igemm_minmax_ukernel_4x4c2__sse2_ld128()
258 const __m128i vprod2x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x13, vnmask2x13), vnmask2x13); in xnn_qs8_igemm_minmax_ukernel_4x4c2__sse2_ld128()
259 const __m128i vprod3x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod3x13, vnmask3x13), vnmask3x13); in xnn_qs8_igemm_minmax_ukernel_4x4c2__sse2_ld128()
D4x4c2-minmax-ssse3-ld64.c236 const __m128i vprod0x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x02, vnmask0x02), vnmask0x02); in xnn_qs8_igemm_minmax_ukernel_4x4c2__ssse3_ld64()
237 const __m128i vprod1x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x02, vnmask1x02), vnmask1x02); in xnn_qs8_igemm_minmax_ukernel_4x4c2__ssse3_ld64()
238 const __m128i vprod2x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x02, vnmask2x02), vnmask2x02); in xnn_qs8_igemm_minmax_ukernel_4x4c2__ssse3_ld64()
239 const __m128i vprod3x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod3x02, vnmask3x02), vnmask3x02); in xnn_qs8_igemm_minmax_ukernel_4x4c2__ssse3_ld64()
256 const __m128i vprod0x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x13, vnmask0x13), vnmask0x13); in xnn_qs8_igemm_minmax_ukernel_4x4c2__ssse3_ld64()
257 const __m128i vprod1x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x13, vnmask1x13), vnmask1x13); in xnn_qs8_igemm_minmax_ukernel_4x4c2__ssse3_ld64()
258 const __m128i vprod2x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x13, vnmask2x13), vnmask2x13); in xnn_qs8_igemm_minmax_ukernel_4x4c2__ssse3_ld64()
259 const __m128i vprod3x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod3x13, vnmask3x13), vnmask3x13); in xnn_qs8_igemm_minmax_ukernel_4x4c2__ssse3_ld64()
D4x4c2-minmax-ssse3-ld128.c236 const __m128i vprod0x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x02, vnmask0x02), vnmask0x02); in xnn_qs8_igemm_minmax_ukernel_4x4c2__ssse3_ld128()
237 const __m128i vprod1x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x02, vnmask1x02), vnmask1x02); in xnn_qs8_igemm_minmax_ukernel_4x4c2__ssse3_ld128()
238 const __m128i vprod2x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x02, vnmask2x02), vnmask2x02); in xnn_qs8_igemm_minmax_ukernel_4x4c2__ssse3_ld128()
239 const __m128i vprod3x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod3x02, vnmask3x02), vnmask3x02); in xnn_qs8_igemm_minmax_ukernel_4x4c2__ssse3_ld128()
256 const __m128i vprod0x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x13, vnmask0x13), vnmask0x13); in xnn_qs8_igemm_minmax_ukernel_4x4c2__ssse3_ld128()
257 const __m128i vprod1x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x13, vnmask1x13), vnmask1x13); in xnn_qs8_igemm_minmax_ukernel_4x4c2__ssse3_ld128()
258 const __m128i vprod2x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x13, vnmask2x13), vnmask2x13); in xnn_qs8_igemm_minmax_ukernel_4x4c2__ssse3_ld128()
259 const __m128i vprod3x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod3x13, vnmask3x13), vnmask3x13); in xnn_qs8_igemm_minmax_ukernel_4x4c2__ssse3_ld128()
D3x4c8-minmax-ssse3-ld64.c162 const __m128i vprod0x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x02, vnmask0x02), vnmask0x02); in xnn_qs8_igemm_minmax_ukernel_3x4c8__ssse3_ld64()
163 const __m128i vprod1x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x02, vnmask1x02), vnmask1x02); in xnn_qs8_igemm_minmax_ukernel_3x4c8__ssse3_ld64()
164 const __m128i vprod2x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x02, vnmask2x02), vnmask2x02); in xnn_qs8_igemm_minmax_ukernel_3x4c8__ssse3_ld64()
178 const __m128i vprod0x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x13, vnmask0x13), vnmask0x13); in xnn_qs8_igemm_minmax_ukernel_3x4c8__ssse3_ld64()
179 const __m128i vprod1x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x13, vnmask1x13), vnmask1x13); in xnn_qs8_igemm_minmax_ukernel_3x4c8__ssse3_ld64()
180 const __m128i vprod2x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x13, vnmask2x13), vnmask2x13); in xnn_qs8_igemm_minmax_ukernel_3x4c8__ssse3_ld64()
D3x4c8-minmax-sse2-ld128.c160 const __m128i vprod0x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x02, vnmask0x02), vnmask0x02); in xnn_qs8_igemm_minmax_ukernel_3x4c8__sse2_ld128()
161 const __m128i vprod1x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x02, vnmask1x02), vnmask1x02); in xnn_qs8_igemm_minmax_ukernel_3x4c8__sse2_ld128()
162 const __m128i vprod2x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x02, vnmask2x02), vnmask2x02); in xnn_qs8_igemm_minmax_ukernel_3x4c8__sse2_ld128()
176 const __m128i vprod0x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x13, vnmask0x13), vnmask0x13); in xnn_qs8_igemm_minmax_ukernel_3x4c8__sse2_ld128()
177 const __m128i vprod1x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x13, vnmask1x13), vnmask1x13); in xnn_qs8_igemm_minmax_ukernel_3x4c8__sse2_ld128()
178 const __m128i vprod2x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x13, vnmask2x13), vnmask2x13); in xnn_qs8_igemm_minmax_ukernel_3x4c8__sse2_ld128()
D3x4c8-minmax-ssse3-ld128.c160 const __m128i vprod0x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x02, vnmask0x02), vnmask0x02); in xnn_qs8_igemm_minmax_ukernel_3x4c8__ssse3_ld128()
161 const __m128i vprod1x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x02, vnmask1x02), vnmask1x02); in xnn_qs8_igemm_minmax_ukernel_3x4c8__ssse3_ld128()
162 const __m128i vprod2x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x02, vnmask2x02), vnmask2x02); in xnn_qs8_igemm_minmax_ukernel_3x4c8__ssse3_ld128()
176 const __m128i vprod0x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x13, vnmask0x13), vnmask0x13); in xnn_qs8_igemm_minmax_ukernel_3x4c8__ssse3_ld128()
177 const __m128i vprod1x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x13, vnmask1x13), vnmask1x13); in xnn_qs8_igemm_minmax_ukernel_3x4c8__ssse3_ld128()
178 const __m128i vprod2x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x13, vnmask2x13), vnmask2x13); in xnn_qs8_igemm_minmax_ukernel_3x4c8__ssse3_ld128()
/external/XNNPACK/src/qu8-igemm/
D4x4c2-minmax-sse2.c206 const __m128i vprod0x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x02, vnmask0x02), vnmask0x02); in xnn_qu8_igemm_minmax_ukernel_4x4c2__sse2()
207 const __m128i vprod1x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x02, vnmask1x02), vnmask1x02); in xnn_qu8_igemm_minmax_ukernel_4x4c2__sse2()
208 const __m128i vprod2x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x02, vnmask2x02), vnmask2x02); in xnn_qu8_igemm_minmax_ukernel_4x4c2__sse2()
209 const __m128i vprod3x02 = _mm_sub_epi64(_mm_xor_si128(vabsprod3x02, vnmask3x02), vnmask3x02); in xnn_qu8_igemm_minmax_ukernel_4x4c2__sse2()
226 const __m128i vprod0x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod0x13, vnmask0x13), vnmask0x13); in xnn_qu8_igemm_minmax_ukernel_4x4c2__sse2()
227 const __m128i vprod1x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod1x13, vnmask1x13), vnmask1x13); in xnn_qu8_igemm_minmax_ukernel_4x4c2__sse2()
228 const __m128i vprod2x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod2x13, vnmask2x13), vnmask2x13); in xnn_qu8_igemm_minmax_ukernel_4x4c2__sse2()
229 const __m128i vprod3x13 = _mm_sub_epi64(_mm_xor_si128(vabsprod3x13, vnmask3x13), vnmask3x13); in xnn_qu8_igemm_minmax_ukernel_4x4c2__sse2()

1234