Searched refs:vb12x1 (Results 1 – 8 of 8) sorted by relevance
/external/XNNPACK/src/qs8-igemm/gen/ |
D | 1x16c8-minmax-neon-mlal-padal.c | 143 … const int8x8_t vb12x1 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof( int8_t)); in xnn_qs8_igemm_minmax_ukernel_1x16c8__neon_mlal_padal() local 145 vprod0x12 = vmlal_s8(vprod0x12, vb12x1, va0x1); in xnn_qs8_igemm_minmax_ukernel_1x16c8__neon_mlal_padal()
|
D | 2x16c8-minmax-neon-mlal-padal.c | 205 … const int8x8_t vb12x1 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof( int8_t)); in xnn_qs8_igemm_minmax_ukernel_2x16c8__neon_mlal_padal() local 208 vprod0x12 = vmlal_s8(vprod0x12, vb12x1, va0x1); in xnn_qs8_igemm_minmax_ukernel_2x16c8__neon_mlal_padal() 209 vprod1x12 = vmlal_s8(vprod1x12, vb12x1, va1x1); in xnn_qs8_igemm_minmax_ukernel_2x16c8__neon_mlal_padal()
|
D | 4x16c8-minmax-neon-mlal-padal.c | 329 … const int8x8_t vb12x1 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof( int8_t)); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal() local 334 vprod0x12 = vmlal_s8(vprod0x12, vb12x1, va0x1); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal() 335 vprod1x12 = vmlal_s8(vprod1x12, vb12x1, va1x1); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal() 336 vprod2x12 = vmlal_s8(vprod2x12, vb12x1, va2x1); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal() 337 vprod3x12 = vmlal_s8(vprod3x12, vb12x1, va3x1); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal()
|
D | 3x16c8-minmax-neon-mlal-padal.c | 267 … const int8x8_t vb12x1 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof( int8_t)); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mlal_padal() local 271 vprod0x12 = vmlal_s8(vprod0x12, vb12x1, va0x1); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mlal_padal() 272 vprod1x12 = vmlal_s8(vprod1x12, vb12x1, va1x1); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mlal_padal() 273 vprod2x12 = vmlal_s8(vprod2x12, vb12x1, va2x1); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mlal_padal()
|
/external/XNNPACK/src/qs8-gemm/gen/ |
D | 1x16c8-minmax-neon-mlal-padal.c | 132 const int8x8_t vb12x1 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof( int8_t)); in xnn_qs8_gemm_minmax_ukernel_1x16c8__neon_mlal_padal() local 134 vprod0x12 = vmlal_s8(vprod0x12, vb12x1, va0x1); in xnn_qs8_gemm_minmax_ukernel_1x16c8__neon_mlal_padal()
|
D | 2x16c8-minmax-neon-mlal-padal.c | 192 const int8x8_t vb12x1 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof( int8_t)); in xnn_qs8_gemm_minmax_ukernel_2x16c8__neon_mlal_padal() local 195 vprod0x12 = vmlal_s8(vprod0x12, vb12x1, va0x1); in xnn_qs8_gemm_minmax_ukernel_2x16c8__neon_mlal_padal() 196 vprod1x12 = vmlal_s8(vprod1x12, vb12x1, va1x1); in xnn_qs8_gemm_minmax_ukernel_2x16c8__neon_mlal_padal()
|
D | 3x16c8-minmax-neon-mlal-padal.c | 252 const int8x8_t vb12x1 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof( int8_t)); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal() local 256 vprod0x12 = vmlal_s8(vprod0x12, vb12x1, va0x1); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal() 257 vprod1x12 = vmlal_s8(vprod1x12, vb12x1, va1x1); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal() 258 vprod2x12 = vmlal_s8(vprod2x12, vb12x1, va2x1); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal()
|
D | 4x16c8-minmax-neon-mlal-padal.c | 312 const int8x8_t vb12x1 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof( int8_t)); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal() local 317 vprod0x12 = vmlal_s8(vprod0x12, vb12x1, va0x1); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal() 318 vprod1x12 = vmlal_s8(vprod1x12, vb12x1, va1x1); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal() 319 vprod2x12 = vmlal_s8(vprod2x12, vb12x1, va2x1); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal() 320 vprod3x12 = vmlal_s8(vprod3x12, vb12x1, va3x1); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal()
|