Searched refs:vb9x0 (Results 1 – 8 of 8) sorted by relevance
/external/XNNPACK/src/qs8-igemm/gen/ |
D | 1x16c8-minmax-neon-mlal-padal.c | 87 const int8x8_t vb9x0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof( int8_t)); in xnn_qs8_igemm_minmax_ukernel_1x16c8__neon_mlal_padal() local 132 int16x8_t vprod0x9 = vmull_s8(vb9x0, va0x0); in xnn_qs8_igemm_minmax_ukernel_1x16c8__neon_mlal_padal()
|
D | 2x16c8-minmax-neon-mlal-padal.c | 113 const int8x8_t vb9x0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof( int8_t)); in xnn_qs8_igemm_minmax_ukernel_2x16c8__neon_mlal_padal() local 185 int16x8_t vprod0x9 = vmull_s8(vb9x0, va0x0); in xnn_qs8_igemm_minmax_ukernel_2x16c8__neon_mlal_padal() 186 int16x8_t vprod1x9 = vmull_s8(vb9x0, va1x0); in xnn_qs8_igemm_minmax_ukernel_2x16c8__neon_mlal_padal()
|
D | 4x16c8-minmax-neon-mlal-padal.c | 165 const int8x8_t vb9x0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof( int8_t)); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal() local 291 int16x8_t vprod0x9 = vmull_s8(vb9x0, va0x0); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal() 292 int16x8_t vprod1x9 = vmull_s8(vb9x0, va1x0); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal() 293 int16x8_t vprod2x9 = vmull_s8(vb9x0, va2x0); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal() 294 int16x8_t vprod3x9 = vmull_s8(vb9x0, va3x0); in xnn_qs8_igemm_minmax_ukernel_4x16c8__neon_mlal_padal()
|
D | 3x16c8-minmax-neon-mlal-padal.c | 139 const int8x8_t vb9x0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof( int8_t)); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mlal_padal() local 238 int16x8_t vprod0x9 = vmull_s8(vb9x0, va0x0); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mlal_padal() 239 int16x8_t vprod1x9 = vmull_s8(vb9x0, va1x0); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mlal_padal() 240 int16x8_t vprod2x9 = vmull_s8(vb9x0, va2x0); in xnn_qs8_igemm_minmax_ukernel_3x16c8__neon_mlal_padal()
|
/external/XNNPACK/src/qs8-gemm/gen/ |
D | 1x16c8-minmax-neon-mlal-padal.c | 76 const int8x8_t vb9x0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof( int8_t)); in xnn_qs8_gemm_minmax_ukernel_1x16c8__neon_mlal_padal() local 121 int16x8_t vprod0x9 = vmull_s8(vb9x0, va0x0); in xnn_qs8_gemm_minmax_ukernel_1x16c8__neon_mlal_padal()
|
D | 2x16c8-minmax-neon-mlal-padal.c | 100 const int8x8_t vb9x0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof( int8_t)); in xnn_qs8_gemm_minmax_ukernel_2x16c8__neon_mlal_padal() local 172 int16x8_t vprod0x9 = vmull_s8(vb9x0, va0x0); in xnn_qs8_gemm_minmax_ukernel_2x16c8__neon_mlal_padal() 173 int16x8_t vprod1x9 = vmull_s8(vb9x0, va1x0); in xnn_qs8_gemm_minmax_ukernel_2x16c8__neon_mlal_padal()
|
D | 3x16c8-minmax-neon-mlal-padal.c | 124 const int8x8_t vb9x0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof( int8_t)); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal() local 223 int16x8_t vprod0x9 = vmull_s8(vb9x0, va0x0); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal() 224 int16x8_t vprod1x9 = vmull_s8(vb9x0, va1x0); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal() 225 int16x8_t vprod2x9 = vmull_s8(vb9x0, va2x0); in xnn_qs8_gemm_minmax_ukernel_3x16c8__neon_mlal_padal()
|
D | 4x16c8-minmax-neon-mlal-padal.c | 148 const int8x8_t vb9x0 = vld1_s8(w); w = (const void*) ((uintptr_t) w + 8 * sizeof( int8_t)); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal() local 274 int16x8_t vprod0x9 = vmull_s8(vb9x0, va0x0); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal() 275 int16x8_t vprod1x9 = vmull_s8(vb9x0, va1x0); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal() 276 int16x8_t vprod2x9 = vmull_s8(vb9x0, va2x0); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal() 277 int16x8_t vprod3x9 = vmull_s8(vb9x0, va3x0); in xnn_qs8_gemm_minmax_ukernel_4x16c8__neon_mlal_padal()
|