Searched refs:vb10x1 (Results 1 – 8 of 8) sorted by relevance
/external/XNNPACK/src/qs8-igemm/gen/ |
D | 3x16c8-minmax-rndnu-neon-mlal.c | 247 … const int8x8_t vb10x1 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof( int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c8__neon_mlal() local 251 vprod0x10 = vmlal_s8(vprod0x10, vb10x1, va0x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c8__neon_mlal() 252 vprod1x10 = vmlal_s8(vprod1x10, vb10x1, va1x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c8__neon_mlal() 253 vprod2x10 = vmlal_s8(vprod2x10, vb10x1, va2x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_3x16c8__neon_mlal()
|
D | 2x16c8-minmax-rndnu-neon-mlal.c | 191 … const int8x8_t vb10x1 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof( int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c8__neon_mlal() local 194 vprod0x10 = vmlal_s8(vprod0x10, vb10x1, va0x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c8__neon_mlal() 195 vprod1x10 = vmlal_s8(vprod1x10, vb10x1, va1x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_2x16c8__neon_mlal()
|
D | 1x16c8-minmax-rndnu-neon-mlal.c | 135 … const int8x8_t vb10x1 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof( int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x16c8__neon_mlal() local 137 vprod0x10 = vmlal_s8(vprod0x10, vb10x1, va0x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_1x16c8__neon_mlal()
|
D | 4x16c8-minmax-rndnu-neon-mlal.c | 303 … const int8x8_t vb10x1 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof( int8_t)); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c8__neon_mlal() local 308 vprod0x10 = vmlal_s8(vprod0x10, vb10x1, va0x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c8__neon_mlal() 309 vprod1x10 = vmlal_s8(vprod1x10, vb10x1, va1x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c8__neon_mlal() 310 vprod2x10 = vmlal_s8(vprod2x10, vb10x1, va2x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c8__neon_mlal() 311 vprod3x10 = vmlal_s8(vprod3x10, vb10x1, va3x1); in xnn_qs8_igemm_minmax_rndnu_ukernel_4x16c8__neon_mlal()
|
/external/XNNPACK/src/qs8-gemm/gen/ |
D | 3x16c8-minmax-rndnu-neon-mlal.c | 232 const int8x8_t vb10x1 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof( int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c8__neon_mlal() local 236 vprod0x10 = vmlal_s8(vprod0x10, vb10x1, va0x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c8__neon_mlal() 237 vprod1x10 = vmlal_s8(vprod1x10, vb10x1, va1x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c8__neon_mlal() 238 vprod2x10 = vmlal_s8(vprod2x10, vb10x1, va2x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_3x16c8__neon_mlal()
|
D | 2x16c8-minmax-rndnu-neon-mlal.c | 178 const int8x8_t vb10x1 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof( int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c8__neon_mlal() local 181 vprod0x10 = vmlal_s8(vprod0x10, vb10x1, va0x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c8__neon_mlal() 182 vprod1x10 = vmlal_s8(vprod1x10, vb10x1, va1x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_2x16c8__neon_mlal()
|
D | 1x16c8-minmax-rndnu-neon-mlal.c | 124 const int8x8_t vb10x1 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof( int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16c8__neon_mlal() local 126 vprod0x10 = vmlal_s8(vprod0x10, vb10x1, va0x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_1x16c8__neon_mlal()
|
D | 4x16c8-minmax-rndnu-neon-mlal.c | 286 const int8x8_t vb10x1 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof( int8_t)); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c8__neon_mlal() local 291 vprod0x10 = vmlal_s8(vprod0x10, vb10x1, va0x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c8__neon_mlal() 292 vprod1x10 = vmlal_s8(vprod1x10, vb10x1, va1x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c8__neon_mlal() 293 vprod2x10 = vmlal_s8(vprod2x10, vb10x1, va2x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c8__neon_mlal() 294 vprod3x10 = vmlal_s8(vprod3x10, vb10x1, va3x1); in xnn_qs8_gemm_minmax_rndnu_ukernel_4x16c8__neon_mlal()
|