/external/XNNPACK/src/f32-gemm/gen/ |
D | 4x2c4-wasmsimd.c | 67 v128_t vacc3x1c4 = vacc0x1c4; in xnn_f32_gemm_ukernel_4x2c4__wasmsimd() local 92 vacc3x1c4 = wasm_f32x4_add(vacc3x1c4, wasm_f32x4_mul(va3, vb1)); in xnn_f32_gemm_ukernel_4x2c4__wasmsimd() 119 vacc3x1c4 = wasm_f32x4_add(vacc3x1c4, wasm_f32x4_mul(wasm_v128_andnot(va3, vmask1), vb1)); in xnn_f32_gemm_ukernel_4x2c4__wasmsimd() 132 wasm_v32x4_shuffle(vacc3x0c4, vacc3x1c4, 0, 4, 1, 5), in xnn_f32_gemm_ukernel_4x2c4__wasmsimd() 133 wasm_v32x4_shuffle(vacc3x0c4, vacc3x1c4, 2, 6, 3, 7)); in xnn_f32_gemm_ukernel_4x2c4__wasmsimd()
|
D | 4x2c4-minmax-wasmsimd-arm.c | 69 v128_t vacc3x1c4 = vacc0x1c4; in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_arm() local 94 vacc3x1c4 = wasm_f32x4_add(vacc3x1c4, wasm_f32x4_mul(va3, vb1)); in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_arm() 121 vacc3x1c4 = wasm_f32x4_add(vacc3x1c4, wasm_f32x4_mul(wasm_v128_andnot(va3, vmask1), vb1)); in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_arm() 134 wasm_v32x4_shuffle(vacc3x0c4, vacc3x1c4, 0, 4, 1, 5), in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_arm() 135 wasm_v32x4_shuffle(vacc3x0c4, vacc3x1c4, 2, 6, 3, 7)); in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_arm()
|
D | 4x2c4-minmax-sse.c | 67 __m128 vacc3x1c4 = vacc0x1c4; in xnn_f32_gemm_minmax_ukernel_4x2c4__sse() local 92 vacc3x1c4 = _mm_add_ps(vacc3x1c4, _mm_mul_ps(va3, vb1)); in xnn_f32_gemm_minmax_ukernel_4x2c4__sse() 118 vacc3x1c4 = _mm_add_ps(vacc3x1c4, _mm_mul_ps(_mm_andnot_ps(vmask1, va3), vb1)); in xnn_f32_gemm_minmax_ukernel_4x2c4__sse() 124 … vacc3x01c2 = _mm_add_ps(_mm_unpacklo_ps(vacc3x0c4, vacc3x1c4), _mm_unpackhi_ps(vacc3x0c4, vacc3x1… in xnn_f32_gemm_minmax_ukernel_4x2c4__sse()
|
D | 4x2c4-relu-wasmsimd.c | 67 v128_t vacc3x1c4 = vacc0x1c4; in xnn_f32_gemm_relu_ukernel_4x2c4__wasmsimd() local 92 vacc3x1c4 = wasm_f32x4_add(vacc3x1c4, wasm_f32x4_mul(va3, vb1)); in xnn_f32_gemm_relu_ukernel_4x2c4__wasmsimd() 119 vacc3x1c4 = wasm_f32x4_add(vacc3x1c4, wasm_f32x4_mul(wasm_v128_andnot(va3, vmask1), vb1)); in xnn_f32_gemm_relu_ukernel_4x2c4__wasmsimd() 132 wasm_v32x4_shuffle(vacc3x0c4, vacc3x1c4, 0, 4, 1, 5), in xnn_f32_gemm_relu_ukernel_4x2c4__wasmsimd() 133 wasm_v32x4_shuffle(vacc3x0c4, vacc3x1c4, 2, 6, 3, 7)); in xnn_f32_gemm_relu_ukernel_4x2c4__wasmsimd()
|
D | 4x2c4-minmax-wasmsimd-x86.c | 67 v128_t vacc3x1c4 = vacc0x1c4; in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_x86() local 92 vacc3x1c4 = wasm_f32x4_add(vacc3x1c4, wasm_f32x4_mul(va3, vb1)); in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_x86() 119 vacc3x1c4 = wasm_f32x4_add(vacc3x1c4, wasm_f32x4_mul(wasm_v128_andnot(va3, vmask1), vb1)); in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_x86() 132 wasm_v32x4_shuffle(vacc3x0c4, vacc3x1c4, 0, 4, 1, 5), in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_x86() 133 wasm_v32x4_shuffle(vacc3x0c4, vacc3x1c4, 2, 6, 3, 7)); in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_x86()
|
/external/XNNPACK/src/f32-igemm/gen/ |
D | 4x2c4-wasmsimd.c | 65 v128_t vacc3x1c4 = vacc0x1c4; in xnn_f32_igemm_ukernel_4x2c4__wasmsimd() local 114 vacc3x1c4 = wasm_f32x4_add(vacc3x1c4, wasm_f32x4_mul(va3, vb1)); in xnn_f32_igemm_ukernel_4x2c4__wasmsimd() 137 vacc3x1c4 = wasm_f32x4_add(vacc3x1c4, wasm_f32x4_mul(wasm_v128_andnot(va3, vmask1), vb1)); in xnn_f32_igemm_ukernel_4x2c4__wasmsimd() 152 wasm_v32x4_shuffle(vacc3x0c4, vacc3x1c4, 0, 4, 1, 5), in xnn_f32_igemm_ukernel_4x2c4__wasmsimd() 153 wasm_v32x4_shuffle(vacc3x0c4, vacc3x1c4, 2, 6, 3, 7)); in xnn_f32_igemm_ukernel_4x2c4__wasmsimd()
|
D | 4x2c4-relu-wasmsimd.c | 65 v128_t vacc3x1c4 = vacc0x1c4; in xnn_f32_igemm_relu_ukernel_4x2c4__wasmsimd() local 114 vacc3x1c4 = wasm_f32x4_add(vacc3x1c4, wasm_f32x4_mul(va3, vb1)); in xnn_f32_igemm_relu_ukernel_4x2c4__wasmsimd() 137 vacc3x1c4 = wasm_f32x4_add(vacc3x1c4, wasm_f32x4_mul(wasm_v128_andnot(va3, vmask1), vb1)); in xnn_f32_igemm_relu_ukernel_4x2c4__wasmsimd() 152 wasm_v32x4_shuffle(vacc3x0c4, vacc3x1c4, 0, 4, 1, 5), in xnn_f32_igemm_relu_ukernel_4x2c4__wasmsimd() 153 wasm_v32x4_shuffle(vacc3x0c4, vacc3x1c4, 2, 6, 3, 7)); in xnn_f32_igemm_relu_ukernel_4x2c4__wasmsimd()
|
D | 4x2c4-minmax-sse.c | 65 __m128 vacc3x1c4 = vacc0x1c4; in xnn_f32_igemm_minmax_ukernel_4x2c4__sse() local 114 vacc3x1c4 = _mm_add_ps(vacc3x1c4, _mm_mul_ps(va3, vb1)); in xnn_f32_igemm_minmax_ukernel_4x2c4__sse() 136 vacc3x1c4 = _mm_add_ps(vacc3x1c4, _mm_mul_ps(_mm_andnot_ps(vmask1, va3), vb1)); in xnn_f32_igemm_minmax_ukernel_4x2c4__sse() 144 … vacc3x01c2 = _mm_add_ps(_mm_unpacklo_ps(vacc3x0c4, vacc3x1c4), _mm_unpackhi_ps(vacc3x0c4, vacc3x1… in xnn_f32_igemm_minmax_ukernel_4x2c4__sse()
|
D | 4x2c4-minmax-wasmsimd-x86.c | 65 v128_t vacc3x1c4 = vacc0x1c4; in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_x86() local 114 vacc3x1c4 = wasm_f32x4_add(vacc3x1c4, wasm_f32x4_mul(va3, vb1)); in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_x86() 137 vacc3x1c4 = wasm_f32x4_add(vacc3x1c4, wasm_f32x4_mul(wasm_v128_andnot(va3, vmask1), vb1)); in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_x86() 152 wasm_v32x4_shuffle(vacc3x0c4, vacc3x1c4, 0, 4, 1, 5), in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_x86() 153 wasm_v32x4_shuffle(vacc3x0c4, vacc3x1c4, 2, 6, 3, 7)); in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_x86()
|
D | 4x2c4-minmax-wasmsimd-arm.c | 67 v128_t vacc3x1c4 = vacc0x1c4; in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_arm() local 116 vacc3x1c4 = wasm_f32x4_add(vacc3x1c4, wasm_f32x4_mul(va3, vb1)); in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_arm() 139 vacc3x1c4 = wasm_f32x4_add(vacc3x1c4, wasm_f32x4_mul(wasm_v128_andnot(va3, vmask1), vb1)); in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_arm() 154 wasm_v32x4_shuffle(vacc3x0c4, vacc3x1c4, 0, 4, 1, 5), in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_arm() 155 wasm_v32x4_shuffle(vacc3x0c4, vacc3x1c4, 2, 6, 3, 7)); in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_arm()
|