/external/XNNPACK/src/f32-gemm/gen/ |
D | 4x2c4-wasmsimd.c | 62 v128_t vacc1x0c4 = vacc0x0c4; in xnn_f32_gemm_ukernel_4x2c4__wasmsimd() local 87 vacc1x0c4 = wasm_f32x4_add(vacc1x0c4, wasm_f32x4_mul(va1, vb0)); in xnn_f32_gemm_ukernel_4x2c4__wasmsimd() 114 vacc1x0c4 = wasm_f32x4_add(vacc1x0c4, wasm_f32x4_mul(wasm_v128_andnot(va1, vmask0), vb0)); in xnn_f32_gemm_ukernel_4x2c4__wasmsimd() 126 wasm_v32x4_shuffle(vacc1x0c4, vacc1x1c4, 0, 4, 1, 5), in xnn_f32_gemm_ukernel_4x2c4__wasmsimd() 127 wasm_v32x4_shuffle(vacc1x0c4, vacc1x1c4, 2, 6, 3, 7)); in xnn_f32_gemm_ukernel_4x2c4__wasmsimd()
|
D | 4x2c4-minmax-wasmsimd-arm.c | 64 v128_t vacc1x0c4 = vacc0x0c4; in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_arm() local 89 vacc1x0c4 = wasm_f32x4_add(vacc1x0c4, wasm_f32x4_mul(va1, vb0)); in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_arm() 116 vacc1x0c4 = wasm_f32x4_add(vacc1x0c4, wasm_f32x4_mul(wasm_v128_andnot(va1, vmask0), vb0)); in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_arm() 128 wasm_v32x4_shuffle(vacc1x0c4, vacc1x1c4, 0, 4, 1, 5), in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_arm() 129 wasm_v32x4_shuffle(vacc1x0c4, vacc1x1c4, 2, 6, 3, 7)); in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_arm()
|
D | 4x2c4-minmax-sse.c | 62 __m128 vacc1x0c4 = vacc0x0c4; in xnn_f32_gemm_minmax_ukernel_4x2c4__sse() local 87 vacc1x0c4 = _mm_add_ps(vacc1x0c4, _mm_mul_ps(va1, vb0)); in xnn_f32_gemm_minmax_ukernel_4x2c4__sse() 113 vacc1x0c4 = _mm_add_ps(vacc1x0c4, _mm_mul_ps(_mm_andnot_ps(vmask0, va1), vb0)); in xnn_f32_gemm_minmax_ukernel_4x2c4__sse() 122 …onst __m128 vacc1x01c2 = _mm_add_ps(_mm_unpacklo_ps(vacc1x0c4, vacc1x1c4), _mm_unpackhi_ps(vacc1x0… in xnn_f32_gemm_minmax_ukernel_4x2c4__sse()
|
D | 4x2c4-relu-wasmsimd.c | 62 v128_t vacc1x0c4 = vacc0x0c4; in xnn_f32_gemm_relu_ukernel_4x2c4__wasmsimd() local 87 vacc1x0c4 = wasm_f32x4_add(vacc1x0c4, wasm_f32x4_mul(va1, vb0)); in xnn_f32_gemm_relu_ukernel_4x2c4__wasmsimd() 114 vacc1x0c4 = wasm_f32x4_add(vacc1x0c4, wasm_f32x4_mul(wasm_v128_andnot(va1, vmask0), vb0)); in xnn_f32_gemm_relu_ukernel_4x2c4__wasmsimd() 126 wasm_v32x4_shuffle(vacc1x0c4, vacc1x1c4, 0, 4, 1, 5), in xnn_f32_gemm_relu_ukernel_4x2c4__wasmsimd() 127 wasm_v32x4_shuffle(vacc1x0c4, vacc1x1c4, 2, 6, 3, 7)); in xnn_f32_gemm_relu_ukernel_4x2c4__wasmsimd()
|
D | 4x2c4-minmax-wasmsimd-x86.c | 62 v128_t vacc1x0c4 = vacc0x0c4; in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_x86() local 87 vacc1x0c4 = wasm_f32x4_add(vacc1x0c4, wasm_f32x4_mul(va1, vb0)); in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_x86() 114 vacc1x0c4 = wasm_f32x4_add(vacc1x0c4, wasm_f32x4_mul(wasm_v128_andnot(va1, vmask0), vb0)); in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_x86() 126 wasm_v32x4_shuffle(vacc1x0c4, vacc1x1c4, 0, 4, 1, 5), in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_x86() 127 wasm_v32x4_shuffle(vacc1x0c4, vacc1x1c4, 2, 6, 3, 7)); in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_x86()
|
/external/XNNPACK/src/f32-igemm/gen/ |
D | 4x2c4-wasmsimd.c | 60 v128_t vacc1x0c4 = vacc0x0c4; in xnn_f32_igemm_ukernel_4x2c4__wasmsimd() local 109 vacc1x0c4 = wasm_f32x4_add(vacc1x0c4, wasm_f32x4_mul(va1, vb0)); in xnn_f32_igemm_ukernel_4x2c4__wasmsimd() 132 vacc1x0c4 = wasm_f32x4_add(vacc1x0c4, wasm_f32x4_mul(wasm_v128_andnot(va1, vmask0), vb0)); in xnn_f32_igemm_ukernel_4x2c4__wasmsimd() 146 wasm_v32x4_shuffle(vacc1x0c4, vacc1x1c4, 0, 4, 1, 5), in xnn_f32_igemm_ukernel_4x2c4__wasmsimd() 147 wasm_v32x4_shuffle(vacc1x0c4, vacc1x1c4, 2, 6, 3, 7)); in xnn_f32_igemm_ukernel_4x2c4__wasmsimd()
|
D | 4x2c4-relu-wasmsimd.c | 60 v128_t vacc1x0c4 = vacc0x0c4; in xnn_f32_igemm_relu_ukernel_4x2c4__wasmsimd() local 109 vacc1x0c4 = wasm_f32x4_add(vacc1x0c4, wasm_f32x4_mul(va1, vb0)); in xnn_f32_igemm_relu_ukernel_4x2c4__wasmsimd() 132 vacc1x0c4 = wasm_f32x4_add(vacc1x0c4, wasm_f32x4_mul(wasm_v128_andnot(va1, vmask0), vb0)); in xnn_f32_igemm_relu_ukernel_4x2c4__wasmsimd() 146 wasm_v32x4_shuffle(vacc1x0c4, vacc1x1c4, 0, 4, 1, 5), in xnn_f32_igemm_relu_ukernel_4x2c4__wasmsimd() 147 wasm_v32x4_shuffle(vacc1x0c4, vacc1x1c4, 2, 6, 3, 7)); in xnn_f32_igemm_relu_ukernel_4x2c4__wasmsimd()
|
D | 4x2c4-minmax-sse.c | 60 __m128 vacc1x0c4 = vacc0x0c4; in xnn_f32_igemm_minmax_ukernel_4x2c4__sse() local 109 vacc1x0c4 = _mm_add_ps(vacc1x0c4, _mm_mul_ps(va1, vb0)); in xnn_f32_igemm_minmax_ukernel_4x2c4__sse() 131 vacc1x0c4 = _mm_add_ps(vacc1x0c4, _mm_mul_ps(_mm_andnot_ps(vmask0, va1), vb0)); in xnn_f32_igemm_minmax_ukernel_4x2c4__sse() 142 …onst __m128 vacc1x01c2 = _mm_add_ps(_mm_unpacklo_ps(vacc1x0c4, vacc1x1c4), _mm_unpackhi_ps(vacc1x0… in xnn_f32_igemm_minmax_ukernel_4x2c4__sse()
|
D | 4x2c4-minmax-wasmsimd-x86.c | 60 v128_t vacc1x0c4 = vacc0x0c4; in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_x86() local 109 vacc1x0c4 = wasm_f32x4_add(vacc1x0c4, wasm_f32x4_mul(va1, vb0)); in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_x86() 132 vacc1x0c4 = wasm_f32x4_add(vacc1x0c4, wasm_f32x4_mul(wasm_v128_andnot(va1, vmask0), vb0)); in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_x86() 146 wasm_v32x4_shuffle(vacc1x0c4, vacc1x1c4, 0, 4, 1, 5), in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_x86() 147 wasm_v32x4_shuffle(vacc1x0c4, vacc1x1c4, 2, 6, 3, 7)); in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_x86()
|
D | 4x2c4-minmax-wasmsimd-arm.c | 62 v128_t vacc1x0c4 = vacc0x0c4; in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_arm() local 111 vacc1x0c4 = wasm_f32x4_add(vacc1x0c4, wasm_f32x4_mul(va1, vb0)); in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_arm() 134 vacc1x0c4 = wasm_f32x4_add(vacc1x0c4, wasm_f32x4_mul(wasm_v128_andnot(va1, vmask0), vb0)); in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_arm() 148 wasm_v32x4_shuffle(vacc1x0c4, vacc1x1c4, 0, 4, 1, 5), in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_arm() 149 wasm_v32x4_shuffle(vacc1x0c4, vacc1x1c4, 2, 6, 3, 7)); in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_arm()
|