| /external/XNNPACK/src/qs8-gemm/gen/ |
| D | 2x8c8-minmax-neon-mlal-padal.c | 100 int16x8_t vprod1x2 = vmull_s8(vb2x0, va1x0); in xnn_qs8_gemm_minmax_ukernel_2x8c8__neon_mlal_padal() local
161 const int16x8_t vprod1x2 = vmull_s8(vb2, va1); in xnn_qs8_gemm_minmax_ukernel_2x8c8__neon_mlal_padal() local
|
| D | 2x4c8-xw-minmax-wasmsimd.c | 89 const v128_t vprod1x2 = wasm_i16x8_mul(vxa1, vxb2); in xnn_qs8_gemm_xw_minmax_ukernel_2x4c8__wasmsimd() local
|
| D | 2x4c8-minmax-wasmsimd-ld128.c | 92 const v128_t vprod1x2 = wasm_i16x8_mul(vxb2, vxa1); in xnn_qs8_gemm_minmax_ukernel_2x4c8__wasmsimd_ld128() local
|
| D | 2x4c8-minmax-wasmsimd-ld64.c | 89 const v128_t vprod1x2 = wasm_i16x8_mul(vxa1, vxb2); in xnn_qs8_gemm_minmax_ukernel_2x4c8__wasmsimd_ld64() local
|
| D | 3x8c8-minmax-neon-mlal-padal.c | 122 int16x8_t vprod1x2 = vmull_s8(vb2x0, va1x0); in xnn_qs8_gemm_minmax_ukernel_3x8c8__neon_mlal_padal() local
206 const int16x8_t vprod1x2 = vmull_s8(vb2, va1); in xnn_qs8_gemm_minmax_ukernel_3x8c8__neon_mlal_padal() local
|
| D | 3x4c8-xw-minmax-wasmsimd.c | 107 const v128_t vprod1x2 = wasm_i16x8_mul(vxa1, vxb2); in xnn_qs8_gemm_xw_minmax_ukernel_3x4c8__wasmsimd() local
|
| D | 3x4c8-minmax-wasmsimd-ld128.c | 110 const v128_t vprod1x2 = wasm_i16x8_mul(vxb2, vxa1); in xnn_qs8_gemm_minmax_ukernel_3x4c8__wasmsimd_ld128() local
|
| D | 3x4c8-minmax-wasmsimd-ld64.c | 107 const v128_t vprod1x2 = wasm_i16x8_mul(vxa1, vxb2); in xnn_qs8_gemm_minmax_ukernel_3x4c8__wasmsimd_ld64() local
|
| D | 2x8c8-minmax-neon-mull-padal.c | 86 const int16x8_t vprod1x2 = vmull_s8(vb2, va1); in xnn_qs8_gemm_minmax_ukernel_2x8c8__neon_mull_padal() local
|
| D | 2x8c16-minmax-neon-mlal-padal.c | 95 int16x8_t vprod1x2 = vmull_s8(vget_low_s8(vb2), vget_low_s8(va1)); in xnn_qs8_gemm_minmax_ukernel_2x8c16__neon_mlal_padal() local
|
| D | 4x8c8-minmax-neon-mlal-padal.c | 144 int16x8_t vprod1x2 = vmull_s8(vb2x0, va1x0); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mlal_padal() local
251 const int16x8_t vprod1x2 = vmull_s8(vb2, va1); in xnn_qs8_gemm_minmax_ukernel_4x8c8__neon_mlal_padal() local
|
| D | 2x16c8-minmax-neon-mlal-padal.c | 124 int16x8_t vprod1x2 = vmull_s8(vb2x0, va1x0); in xnn_qs8_gemm_minmax_ukernel_2x16c8__neon_mlal_padal() local
241 const int16x8_t vprod1x2 = vmull_s8(vb2, va1); in xnn_qs8_gemm_minmax_ukernel_2x16c8__neon_mlal_padal() local
|
| D | 3x8c16-minmax-neon-mlal-padal.c | 116 int16x8_t vprod1x2 = vmull_s8(vget_low_s8(vb2), vget_low_s8(va1)); in xnn_qs8_gemm_minmax_ukernel_3x8c16__neon_mlal_padal() local
|
| /external/XNNPACK/src/qs8-igemm/gen/ |
| D | 2x8c8-minmax-neon-mlal-padal.c | 113 int16x8_t vprod1x2 = vmull_s8(vb2x0, va1x0); in xnn_qs8_igemm_minmax_ukernel_2x8c8__neon_mlal_padal() local
174 const int16x8_t vprod1x2 = vmull_s8(vb2, va1); in xnn_qs8_igemm_minmax_ukernel_2x8c8__neon_mlal_padal() local
|
| D | 2x4c8-minmax-wasmsimd-ld128.c | 105 const v128_t vprod1x2 = wasm_i16x8_mul(vxb2, vxa1); in xnn_qs8_igemm_minmax_ukernel_2x4c8__wasmsimd_ld128() local
|
| D | 2x4c8-minmax-wasmsimd-ld64.c | 102 const v128_t vprod1x2 = wasm_i16x8_mul(vxa1, vxb2); in xnn_qs8_igemm_minmax_ukernel_2x4c8__wasmsimd_ld64() local
|
| D | 3x8c8-minmax-neon-mlal-padal.c | 137 int16x8_t vprod1x2 = vmull_s8(vb2x0, va1x0); in xnn_qs8_igemm_minmax_ukernel_3x8c8__neon_mlal_padal() local
221 const int16x8_t vprod1x2 = vmull_s8(vb2, va1); in xnn_qs8_igemm_minmax_ukernel_3x8c8__neon_mlal_padal() local
|
| D | 3x4c8-minmax-wasmsimd-ld64.c | 122 const v128_t vprod1x2 = wasm_i16x8_mul(vxa1, vxb2); in xnn_qs8_igemm_minmax_ukernel_3x4c8__wasmsimd_ld64() local
|
| D | 3x4c8-minmax-wasmsimd-ld128.c | 125 const v128_t vprod1x2 = wasm_i16x8_mul(vxb2, vxa1); in xnn_qs8_igemm_minmax_ukernel_3x4c8__wasmsimd_ld128() local
|
| D | 4x8c8-minmax-neon-mlal-padal.c | 161 int16x8_t vprod1x2 = vmull_s8(vb2x0, va1x0); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mlal_padal() local
268 const int16x8_t vprod1x2 = vmull_s8(vb2, va1); in xnn_qs8_igemm_minmax_ukernel_4x8c8__neon_mlal_padal() local
|
| D | 2x8c16-minmax-neon-mlal-padal.c | 108 int16x8_t vprod1x2 = vmull_s8(vget_low_s8(vb2), vget_low_s8(va1)); in xnn_qs8_igemm_minmax_ukernel_2x8c16__neon_mlal_padal() local
|
| D | 2x8c8-minmax-neon-mull-padal.c | 99 const int16x8_t vprod1x2 = vmull_s8(vb2, va1); in xnn_qs8_igemm_minmax_ukernel_2x8c8__neon_mull_padal() local
|
| D | 2x16c8-minmax-neon-mlal-padal.c | 137 int16x8_t vprod1x2 = vmull_s8(vb2x0, va1x0); in xnn_qs8_igemm_minmax_ukernel_2x16c8__neon_mlal_padal() local
254 const int16x8_t vprod1x2 = vmull_s8(vb2, va1); in xnn_qs8_igemm_minmax_ukernel_2x16c8__neon_mlal_padal() local
|
| D | 3x8c8-minmax-neon-mull-padal.c | 120 const int16x8_t vprod1x2 = vmull_s8(vb2, va1); in xnn_qs8_igemm_minmax_ukernel_3x8c8__neon_mull_padal() local
|
| D | 3x8c16-minmax-neon-mlal-padal.c | 131 int16x8_t vprod1x2 = vmull_s8(vget_low_s8(vb2), vget_low_s8(va1)); in xnn_qs8_igemm_minmax_ukernel_3x8c16__neon_mlal_padal() local
|