| /aosp_15_r20/external/XNNPACK/src/f32-gemm/gen/ |
| H A D | 6x2-minmax-neon-lane-ld64.c | 87 const float32x2_t va5 = vld1_f32(a5); a5 += 2; in xnn_f32_gemm_minmax_ukernel_6x2__neon_lane_ld64() local
112 const float32x2_t va5 = vld1_dup_f32(a5); a5 += 1; in xnn_f32_gemm_minmax_ukernel_6x2__neon_lane_ld64() local
|
| H A D | 6x2-minmax-neonfma-lane-ld64.c | 87 const float32x2_t va5 = vld1_f32(a5); a5 += 2; in xnn_f32_gemm_minmax_ukernel_6x2__neonfma_lane_ld64() local
142 const float32x2_t va5 = vld1_dup_f32(a5); a5 += 1; in xnn_f32_gemm_minmax_ukernel_6x2__neonfma_lane_ld64() local
|
| H A D | 6x8-minmax-neon-lane-ld64.c | 93 const float32x2_t va5 = vld1_f32(a5); a5 += 2; in xnn_f32_gemm_minmax_ukernel_6x8__neon_lane_ld64() local
132 const float32x4_t va5 = vld1q_dup_f32(a5); a5 += 1; in xnn_f32_gemm_minmax_ukernel_6x8__neon_lane_ld64() local
|
| H A D | 6x8-minmax-neonfma-lane-ld64.c | 93 const float32x2_t va5 = vld1_f32(a5); a5 += 2; in xnn_f32_gemm_minmax_ukernel_6x8__neonfma_lane_ld64() local
132 const float32x4_t va5 = vld1q_dup_f32(a5); a5 += 1; in xnn_f32_gemm_minmax_ukernel_6x8__neonfma_lane_ld64() local
|
| H A D | 6x8s4-wasmrelaxedsimd-fma.c | 98 v128_t va5 = wasm_v128_load(a5); in xnn_f32_gemm_ukernel_6x8s4__wasmrelaxedsimd_fma() local
202 v128_t va5 = wasm_v128_load(a5); in xnn_f32_gemm_ukernel_6x8s4__wasmrelaxedsimd_fma() local
|
| H A D | 6x8-minmax-neonfma-lane-ld128.c | 93 const float32x4_t va5 = vld1q_f32(a5); a5 += 4; in xnn_f32_gemm_minmax_ukernel_6x8__neonfma_lane_ld128() local
167 const float32x4_t va5 = vld1q_dup_f32(a5); a5 += 1; in xnn_f32_gemm_minmax_ukernel_6x8__neonfma_lane_ld128() local
|
| H A D | 6x8s4-relu-wasmrelaxedsimd-fma.c | 98 v128_t va5 = wasm_v128_load(a5); in xnn_f32_gemm_relu_ukernel_6x8s4__wasmrelaxedsimd_fma() local
202 v128_t va5 = wasm_v128_load(a5); in xnn_f32_gemm_relu_ukernel_6x8s4__wasmrelaxedsimd_fma() local
|
| H A D | 6x8s4-minmax-wasmsimd-arm.c | 100 v128_t va5 = wasm_v128_load(a5); in xnn_f32_gemm_minmax_ukernel_6x8s4__wasmsimd_arm() local
204 v128_t va5 = wasm_v128_load(a5); in xnn_f32_gemm_minmax_ukernel_6x8s4__wasmsimd_arm() local
|
| H A D | 6x8s4-minmax-wasmsimd-x86.c | 100 v128_t va5 = wasm_v128_load(a5); in xnn_f32_gemm_minmax_ukernel_6x8s4__wasmsimd_x86() local
204 v128_t va5 = wasm_v128_load(a5); in xnn_f32_gemm_minmax_ukernel_6x8s4__wasmsimd_x86() local
|
| H A D | 6x8s4-minmax-wasmrelaxedsimd.c | 100 v128_t va5 = wasm_v128_load(a5); in xnn_f32_gemm_minmax_ukernel_6x8s4__wasmrelaxedsimd() local
204 v128_t va5 = wasm_v128_load(a5); in xnn_f32_gemm_minmax_ukernel_6x8s4__wasmrelaxedsimd() local
|
| H A D | 6x8s4-wasmsimd.c | 98 v128_t va5 = wasm_v128_load(a5); in xnn_f32_gemm_ukernel_6x8s4__wasmsimd() local
202 v128_t va5 = wasm_v128_load(a5); in xnn_f32_gemm_ukernel_6x8s4__wasmsimd() local
|
| H A D | 6x8-minmax-neon-lane-ld128.c | 93 const float32x4_t va5 = vld1q_f32(a5); a5 += 4; in xnn_f32_gemm_minmax_ukernel_6x8__neon_lane_ld128() local
167 const float32x4_t va5 = vld1q_dup_f32(a5); a5 += 1; in xnn_f32_gemm_minmax_ukernel_6x8__neon_lane_ld128() local
|
| H A D | 6x8s4-minmax-wasmrelaxedsimd-fma.c | 100 v128_t va5 = wasm_v128_load(a5); in xnn_f32_gemm_minmax_ukernel_6x8s4__wasmrelaxedsimd_fma() local
204 v128_t va5 = wasm_v128_load(a5); in xnn_f32_gemm_minmax_ukernel_6x8s4__wasmrelaxedsimd_fma() local
|
| H A D | 6x8s4-relu-wasmsimd.c | 98 v128_t va5 = wasm_v128_load(a5); in xnn_f32_gemm_relu_ukernel_6x8s4__wasmsimd() local
202 v128_t va5 = wasm_v128_load(a5); in xnn_f32_gemm_relu_ukernel_6x8s4__wasmsimd() local
|
| /aosp_15_r20/external/XNNPACK/src/f32-igemm/gen/ |
| H A D | 6x2-minmax-neon-lane-ld64.c | 115 const float32x2_t va5 = vld1_f32(a5); a5 += 2; in xnn_f32_igemm_minmax_ukernel_6x2__neon_lane_ld64() local
140 const float32x2_t va5 = vld1_dup_f32(a5); in xnn_f32_igemm_minmax_ukernel_6x2__neon_lane_ld64() local
|
| H A D | 6x2-minmax-neonfma-lane-ld64.c | 115 const float32x2_t va5 = vld1_f32(a5); a5 += 2; in xnn_f32_igemm_minmax_ukernel_6x2__neonfma_lane_ld64() local
170 const float32x2_t va5 = vld1_dup_f32(a5); in xnn_f32_igemm_minmax_ukernel_6x2__neonfma_lane_ld64() local
|
| H A D | 6x8-minmax-neonfma-lane-ld64.c | 121 const float32x2_t va5 = vld1_f32(a5); a5 += 2; in xnn_f32_igemm_minmax_ukernel_6x8__neonfma_lane_ld64() local
161 const float32x4_t va5 = vld1q_dup_f32(a5); in xnn_f32_igemm_minmax_ukernel_6x8__neonfma_lane_ld64() local
|
| H A D | 6x8-minmax-neon-lane-ld64.c | 121 const float32x2_t va5 = vld1_f32(a5); a5 += 2; in xnn_f32_igemm_minmax_ukernel_6x8__neon_lane_ld64() local
161 const float32x4_t va5 = vld1q_dup_f32(a5); in xnn_f32_igemm_minmax_ukernel_6x8__neon_lane_ld64() local
|
| H A D | 6x8s4-wasmrelaxedsimd-fma.c | 126 v128_t va5 = wasm_v128_load(a5); in xnn_f32_igemm_ukernel_6x8s4__wasmrelaxedsimd_fma() local
230 v128_t va5 = wasm_v128_load(a5); in xnn_f32_igemm_ukernel_6x8s4__wasmrelaxedsimd_fma() local
|
| /aosp_15_r20/external/pytorch/aten/src/ATen/native/quantized/cpu/qnnpack/src/sgemm/ |
| H A D | 6x8-neon.c | 72 const float32x2_t va5 = vld1_f32(a5); in pytorch_sgemm_ukernel_6x8__neon() local
151 const float32x4_t va5 = vld1q_dup_f32(a5); in pytorch_sgemm_ukernel_6x8__neon() local
|
| /aosp_15_r20/external/pytorch/aten/src/ATen/native/quantized/cpu/qnnpack/src/hgemm/ |
| H A D | 8x8-neonfp16arith.c | 75 const float16x4_t va5 = vld1_f16(a5); in pytorch_hgemm_ukernel_8x8__neonfp16arith() local
151 const float16x4_t va5 = vreinterpret_f16_u64(vshl_u64( in pytorch_hgemm_ukernel_8x8__neonfp16arith() local
|
| /aosp_15_r20/external/XNNPACK/src/f32-gemm/gen-inc/ |
| H A D | 6x8inc-minmax-neonfma-lane-ld64.c | 95 const float32x2_t va5 = vld1_f32(a5); a5 += 2; in xnn_f32_gemminc_minmax_ukernel_6x8__neonfma_lane_ld64() local
134 const float32x4_t va5 = vld1q_dup_f32(a5); a5 += 1; in xnn_f32_gemminc_minmax_ukernel_6x8__neonfma_lane_ld64() local
|
| H A D | 6x8inc-minmax-neon-lane-ld64.c | 95 const float32x2_t va5 = vld1_f32(a5); a5 += 2; in xnn_f32_gemminc_minmax_ukernel_6x8__neon_lane_ld64() local
134 const float32x4_t va5 = vld1q_dup_f32(a5); a5 += 1; in xnn_f32_gemminc_minmax_ukernel_6x8__neon_lane_ld64() local
|
| H A D | 6x8inc-minmax-neon-lane-ld128.c | 95 const float32x4_t va5 = vld1q_f32(a5); a5 += 4; in xnn_f32_gemminc_minmax_ukernel_6x8__neon_lane_ld128() local
169 const float32x4_t va5 = vld1q_dup_f32(a5); a5 += 1; in xnn_f32_gemminc_minmax_ukernel_6x8__neon_lane_ld128() local
|
| H A D | 6x8inc-minmax-neonfma-lane-ld128.c | 95 const float32x4_t va5 = vld1q_f32(a5); a5 += 4; in xnn_f32_gemminc_minmax_ukernel_6x8__neonfma_lane_ld128() local
169 const float32x4_t va5 = vld1q_dup_f32(a5); a5 += 1; in xnn_f32_gemminc_minmax_ukernel_6x8__neonfma_lane_ld128() local
|