| /external/XNNPACK/src/qs8-gavgpool/gen/ |
| D | 7x-minmax-wasmsimd-c24-acc2.c | 125 const v128_t vabsaccKLMN = wasm_i32x4_abs(vaccKLMN); in xnn_qs8_gavgpool_minmax_ukernel_7x__wasmsimd_c24_acc2() local
132 const v128_t vsgnaccKLMN = wasm_i32x4_gt(vabsaccKLMN, vaccKLMN); in xnn_qs8_gavgpool_minmax_ukernel_7x__wasmsimd_c24_acc2()
144 const v128_t vabsaccKL = wasm_v32x4_shuffle(vabsaccKLMN, vzero, 0, 4, 1, 5); in xnn_qs8_gavgpool_minmax_ukernel_7x__wasmsimd_c24_acc2()
145 const v128_t vabsaccMN = wasm_v32x4_shuffle(vabsaccKLMN, vzero, 2, 6, 3, 7); in xnn_qs8_gavgpool_minmax_ukernel_7x__wasmsimd_c24_acc2()
|
| D | 7x-minmax-sse41-c24-acc2.c | 125 const __m128i vabsaccKLMN = _mm_abs_epi32(vaccKLMN); in xnn_qs8_gavgpool_minmax_ukernel_7x__sse41_c24_acc2() local
132 const __m128i vabsaccLN = _mm_shuffle_epi32(vabsaccKLMN, _MM_SHUFFLE(3, 3, 1, 1)); in xnn_qs8_gavgpool_minmax_ukernel_7x__sse41_c24_acc2()
144 const __m128i vabsprodKM = _mm_mul_epu32(vabsaccKLMN, vmultiplier); in xnn_qs8_gavgpool_minmax_ukernel_7x__sse41_c24_acc2()
|
| D | 7x-minmax-ssse3-c24-acc2.c | 149 const __m128i vabsaccKLMN = _mm_abs_epi32(vaccKLMN); in xnn_qs8_gavgpool_minmax_ukernel_7x__ssse3_c24_acc2() local
156 const __m128i vabsaccLN = _mm_shuffle_epi32(vabsaccKLMN, _MM_SHUFFLE(3, 3, 1, 1)); in xnn_qs8_gavgpool_minmax_ukernel_7x__ssse3_c24_acc2()
168 const __m128i vabsprodKM = _mm_mul_epu32(vabsaccKLMN, vmultiplier); in xnn_qs8_gavgpool_minmax_ukernel_7x__ssse3_c24_acc2()
|
| D | 7x-minmax-sse2-c24-acc2.c | 156 const __m128i vabsaccKLMN = _mm_sub_epi32(_mm_xor_si128(vaccKLMN, vsgnaccKLMN), vsgnaccKLMN); in xnn_qs8_gavgpool_minmax_ukernel_7x__sse2_c24_acc2() local
163 const __m128i vabsaccLN = _mm_shuffle_epi32(vabsaccKLMN, _MM_SHUFFLE(3, 3, 1, 1)); in xnn_qs8_gavgpool_minmax_ukernel_7x__sse2_c24_acc2()
175 const __m128i vabsprodKM = _mm_mul_epu32(vabsaccKLMN, vmultiplier); in xnn_qs8_gavgpool_minmax_ukernel_7x__sse2_c24_acc2()
|
| D | 7p7x-minmax-wasmsimd-c24-acc2.c | 360 const v128_t vabsaccKLMN = wasm_i32x4_abs(vaccKLMN); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__wasmsimd_c24_acc2() local
367 const v128_t vsgnaccKLMN = wasm_i32x4_gt(vabsaccKLMN, vaccKLMN); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__wasmsimd_c24_acc2()
379 const v128_t vabsaccKL = wasm_v32x4_shuffle(vabsaccKLMN, vzero, 0, 4, 1, 5); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__wasmsimd_c24_acc2()
380 const v128_t vabsaccMN = wasm_v32x4_shuffle(vabsaccKLMN, vzero, 2, 6, 3, 7); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__wasmsimd_c24_acc2()
|
| D | 7p7x-minmax-sse41-c24-acc2.c | 364 const __m128i vabsaccKLMN = _mm_abs_epi32(vaccKLMN); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__sse41_c24_acc2() local
371 const __m128i vabsaccLN = _mm_shuffle_epi32(vabsaccKLMN, _MM_SHUFFLE(3, 3, 1, 1)); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__sse41_c24_acc2()
383 const __m128i vabsprodKM = _mm_mul_epu32(vabsaccKLMN, vmultiplier); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__sse41_c24_acc2()
|
| D | 7p7x-minmax-sse2-c24-acc2.c | 459 const __m128i vabsaccKLMN = _mm_sub_epi32(_mm_xor_si128(vaccKLMN, vsgnaccKLMN), vsgnaccKLMN); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__sse2_c24_acc2() local
466 const __m128i vabsaccLN = _mm_shuffle_epi32(vabsaccKLMN, _MM_SHUFFLE(3, 3, 1, 1)); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__sse2_c24_acc2()
478 const __m128i vabsprodKM = _mm_mul_epu32(vabsaccKLMN, vmultiplier); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__sse2_c24_acc2()
|
| D | 7p7x-minmax-ssse3-c24-acc2.c | 452 const __m128i vabsaccKLMN = _mm_abs_epi32(vaccKLMN); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__ssse3_c24_acc2() local
459 const __m128i vabsaccLN = _mm_shuffle_epi32(vabsaccKLMN, _MM_SHUFFLE(3, 3, 1, 1)); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__ssse3_c24_acc2()
471 const __m128i vabsprodKM = _mm_mul_epu32(vabsaccKLMN, vmultiplier); in xnn_qs8_gavgpool_minmax_ukernel_7p7x__ssse3_c24_acc2()
|
| /external/XNNPACK/src/qs8-dwconv/gen/ |
| D | up24x9-minmax-ssse3-mul16.c | 369 const __m128i vabsaccKLMN = _mm_abs_epi32(vaccKLMN); in xnn_qs8_dwconv_minmax_ukernel_up24x9__ssse3_mul16() local
386 const __m128i vabsaccLN = _mm_shuffle_epi32(vabsaccKLMN, _MM_SHUFFLE(3, 3, 1, 1)); in xnn_qs8_dwconv_minmax_ukernel_up24x9__ssse3_mul16()
387 const __m128i vabsprodKM = _mm_mul_epu32(vabsaccKLMN, vmultiplier); in xnn_qs8_dwconv_minmax_ukernel_up24x9__ssse3_mul16()
|
| D | up24x9-minmax-sse2-mul16.c | 369 const __m128i vabsaccKLMN = _mm_sub_epi32(_mm_xor_si128(vaccKLMN, vnmaskKLMN), vnmaskKLMN); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse2_mul16() local
386 const __m128i vabsaccLN = _mm_shuffle_epi32(vabsaccKLMN, _MM_SHUFFLE(3, 3, 1, 1)); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse2_mul16()
387 const __m128i vabsprodKM = _mm_mul_epu32(vabsaccKLMN, vmultiplier); in xnn_qs8_dwconv_minmax_ukernel_up24x9__sse2_mul16()
|