/external/XNNPACK/src/qs8-gavgpool/gen/ |
D | 7p7x-minmax-wasmsimd-c24-acc2.c | 58 const v128_t vxi3xGHIJKLMN = wasm_i16x8_load_8x8(i3 + 16); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__wasmsimd_c24_acc2() local 78 v128_t vacc1xGHIJKLMN = wasm_i16x8_add(vxi2xGHIJKLMN, vxi3xGHIJKLMN); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__wasmsimd_c24_acc2() 174 const v128_t vxi3xGHIJKLMN = wasm_i16x8_load_8x8(i3 + 16); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__wasmsimd_c24_acc2() local 194 v128_t vacc1xGHIJKLMN = wasm_i16x8_add(vxi2xGHIJKLMN, vxi3xGHIJKLMN); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__wasmsimd_c24_acc2() 310 const v128_t vxi3xGHIJKLMN = wasm_i16x8_load_8x8(i3 + 16); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__wasmsimd_c24_acc2() local 330 v128_t vacc1xGHIJKLMN = wasm_i16x8_add(vxi2xGHIJKLMN, vxi3xGHIJKLMN); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__wasmsimd_c24_acc2()
|
D | 7p7x-minmax-sse41-c24-acc2.c | 58 const __m128i vxi3xGHIJKLMN = _mm_cvtepi8_epi16(_mm_loadl_epi64((const __m128i*) (i3 + 16))); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__sse41_c24_acc2() local 79 __m128i vacc1xGHIJKLMN = _mm_add_epi16(vxi2xGHIJKLMN, vxi3xGHIJKLMN); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__sse41_c24_acc2() 176 const __m128i vxi3xGHIJKLMN = _mm_cvtepi8_epi16(_mm_loadl_epi64((const __m128i*) (i3 + 16))); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__sse41_c24_acc2() local 197 __m128i vacc1xGHIJKLMN = _mm_add_epi16(vxi2xGHIJKLMN, vxi3xGHIJKLMN); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__sse41_c24_acc2() 313 const __m128i vxi3xGHIJKLMN = _mm_cvtepi8_epi16(_mm_loadl_epi64((const __m128i*) (i3 + 16))); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__sse41_c24_acc2() local 334 __m128i vacc1xGHIJKLMN = _mm_add_epi16(vxi2xGHIJKLMN, vxi3xGHIJKLMN); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__sse41_c24_acc2()
|
D | 7p7x-minmax-sse2-c24-acc2.c | 84 …const __m128i vxi3xGHIJKLMN = _mm_unpacklo_epi8(vi3xGHIJKLMN, _mm_cmpgt_epi8(_mm_setzero_si128(), … in xnn_qs8_gavgpool_minmax_ukernel_7p7x__sse2_c24_acc2() local 100 __m128i vacc1xGHIJKLMN = _mm_add_epi16(vxi2xGHIJKLMN, vxi3xGHIJKLMN); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__sse2_c24_acc2() 234 …const __m128i vxi3xGHIJKLMN = _mm_unpacklo_epi8(vi3xGHIJKLMN, _mm_cmpgt_epi8(_mm_setzero_si128(), … in xnn_qs8_gavgpool_minmax_ukernel_7p7x__sse2_c24_acc2() local 250 __m128i vacc1xGHIJKLMN = _mm_add_epi16(vxi2xGHIJKLMN, vxi3xGHIJKLMN); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__sse2_c24_acc2() 403 …const __m128i vxi3xGHIJKLMN = _mm_unpacklo_epi8(vi3xGHIJKLMN, _mm_cmpgt_epi8(_mm_setzero_si128(), … in xnn_qs8_gavgpool_minmax_ukernel_7p7x__sse2_c24_acc2() local 419 __m128i vacc1xGHIJKLMN = _mm_add_epi16(vxi2xGHIJKLMN, vxi3xGHIJKLMN); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__sse2_c24_acc2()
|
D | 7p7x-minmax-ssse3-c24-acc2.c | 84 …const __m128i vxi3xGHIJKLMN = _mm_unpacklo_epi8(vi3xGHIJKLMN, _mm_cmpgt_epi8(_mm_setzero_si128(), … in xnn_qs8_gavgpool_minmax_ukernel_7p7x__ssse3_c24_acc2() local 100 __m128i vacc1xGHIJKLMN = _mm_add_epi16(vxi2xGHIJKLMN, vxi3xGHIJKLMN); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__ssse3_c24_acc2() 234 …const __m128i vxi3xGHIJKLMN = _mm_unpacklo_epi8(vi3xGHIJKLMN, _mm_cmpgt_epi8(_mm_setzero_si128(), … in xnn_qs8_gavgpool_minmax_ukernel_7p7x__ssse3_c24_acc2() local 250 __m128i vacc1xGHIJKLMN = _mm_add_epi16(vxi2xGHIJKLMN, vxi3xGHIJKLMN); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__ssse3_c24_acc2() 403 …const __m128i vxi3xGHIJKLMN = _mm_unpacklo_epi8(vi3xGHIJKLMN, _mm_cmpgt_epi8(_mm_setzero_si128(), … in xnn_qs8_gavgpool_minmax_ukernel_7p7x__ssse3_c24_acc2() local 419 __m128i vacc1xGHIJKLMN = _mm_add_epi16(vxi2xGHIJKLMN, vxi3xGHIJKLMN); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__ssse3_c24_acc2()
|
D | 7x-minmax-sse41-c24-acc2.c | 75 const __m128i vxi3xGHIJKLMN = _mm_cvtepi8_epi16(_mm_loadl_epi64((const __m128i*) (i3 + 16))); in xnn_qs8_gavgpool_minmax_ukernel_7x__sse41_c24_acc2() local 96 __m128i vacc1xGHIJKLMN = _mm_add_epi16(vxi2xGHIJKLMN, vxi3xGHIJKLMN); in xnn_qs8_gavgpool_minmax_ukernel_7x__sse41_c24_acc2()
|
D | 7x-minmax-wasmsimd-c24-acc2.c | 76 const v128_t vxi3xGHIJKLMN = wasm_i16x8_load_8x8(i3 + 16); in xnn_qs8_gavgpool_minmax_ukernel_7x__wasmsimd_c24_acc2() local 96 v128_t vacc1xGHIJKLMN = wasm_i16x8_add(vxi2xGHIJKLMN, vxi3xGHIJKLMN); in xnn_qs8_gavgpool_minmax_ukernel_7x__wasmsimd_c24_acc2()
|
D | 7x-minmax-sse2-c24-acc2.c | 101 …const __m128i vxi3xGHIJKLMN = _mm_unpacklo_epi8(vi3xGHIJKLMN, _mm_cmpgt_epi8(_mm_setzero_si128(), … in xnn_qs8_gavgpool_minmax_ukernel_7x__sse2_c24_acc2() local 117 __m128i vacc1xGHIJKLMN = _mm_add_epi16(vxi2xGHIJKLMN, vxi3xGHIJKLMN); in xnn_qs8_gavgpool_minmax_ukernel_7x__sse2_c24_acc2()
|
D | 7x-minmax-ssse3-c24-acc2.c | 101 …const __m128i vxi3xGHIJKLMN = _mm_unpacklo_epi8(vi3xGHIJKLMN, _mm_cmpgt_epi8(_mm_setzero_si128(), … in xnn_qs8_gavgpool_minmax_ukernel_7x__ssse3_c24_acc2() local 117 __m128i vacc1xGHIJKLMN = _mm_add_epi16(vxi2xGHIJKLMN, vxi3xGHIJKLMN); in xnn_qs8_gavgpool_minmax_ukernel_7x__ssse3_c24_acc2()
|
/external/XNNPACK/src/qs8-dwconv/gen/ |
D | up24x9-minmax-sse41-mul16.c | 187 const __m128i vxi3xGHIJKLMN = _mm_cvtepi8_epi16(vi3xGHIJKLMN); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse41_mul16() local 197 const __m128i vp3xGHIJKLMNlo = _mm_mullo_epi16(vxi3xGHIJKLMN, vxk3xGHIJKLMN); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse41_mul16() 198 const __m128i vp3xGHIJKLMNhi = _mm_mulhi_epi16(vxi3xGHIJKLMN, vxk3xGHIJKLMN); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse41_mul16()
|
D | up24x9-minmax-ssse3-mul16.c | 190 …const __m128i vxi3xGHIJKLMN = _mm_unpacklo_epi8(vi3xGHIJKLMN, _mm_cmpgt_epi8(_mm_setzero_si128(), … in xnn_qs8_dwconv_minmax_ukernel_up24x9__ssse3_mul16() local 197 const __m128i vp3xGHIJKLMNlo = _mm_mullo_epi16(vxi3xGHIJKLMN, vxk3xGHIJKLMN); in xnn_qs8_dwconv_minmax_ukernel_up24x9__ssse3_mul16() 198 const __m128i vp3xGHIJKLMNhi = _mm_mulhi_epi16(vxi3xGHIJKLMN, vxk3xGHIJKLMN); in xnn_qs8_dwconv_minmax_ukernel_up24x9__ssse3_mul16()
|
D | up24x9-minmax-sse2-mul16.c | 190 …const __m128i vxi3xGHIJKLMN = _mm_unpacklo_epi8(vi3xGHIJKLMN, _mm_cmpgt_epi8(_mm_setzero_si128(), … in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse2_mul16() local 197 const __m128i vp3xGHIJKLMNlo = _mm_mullo_epi16(vxi3xGHIJKLMN, vxk3xGHIJKLMN); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse2_mul16() 198 const __m128i vp3xGHIJKLMNhi = _mm_mulhi_epi16(vxi3xGHIJKLMN, vxk3xGHIJKLMN); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse2_mul16()
|