/external/XNNPACK/src/f32-gemm/gen/ |
D | 4x2c4-wasmsimd.c | 64 v128_t vacc2x0c4 = vacc0x0c4; in xnn_f32_gemm_ukernel_4x2c4__wasmsimd() local 89 vacc2x0c4 = wasm_f32x4_add(vacc2x0c4, wasm_f32x4_mul(va2, vb0)); in xnn_f32_gemm_ukernel_4x2c4__wasmsimd() 116 vacc2x0c4 = wasm_f32x4_add(vacc2x0c4, wasm_f32x4_mul(wasm_v128_andnot(va2, vmask0), vb0)); in xnn_f32_gemm_ukernel_4x2c4__wasmsimd() 129 wasm_v32x4_shuffle(vacc2x0c4, vacc2x1c4, 0, 4, 1, 5), in xnn_f32_gemm_ukernel_4x2c4__wasmsimd() 130 wasm_v32x4_shuffle(vacc2x0c4, vacc2x1c4, 2, 6, 3, 7)); in xnn_f32_gemm_ukernel_4x2c4__wasmsimd()
|
D | 4x2c4-minmax-wasmsimd-arm.c | 66 v128_t vacc2x0c4 = vacc0x0c4; in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_arm() local 91 vacc2x0c4 = wasm_f32x4_add(vacc2x0c4, wasm_f32x4_mul(va2, vb0)); in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_arm() 118 vacc2x0c4 = wasm_f32x4_add(vacc2x0c4, wasm_f32x4_mul(wasm_v128_andnot(va2, vmask0), vb0)); in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_arm() 131 wasm_v32x4_shuffle(vacc2x0c4, vacc2x1c4, 0, 4, 1, 5), in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_arm() 132 wasm_v32x4_shuffle(vacc2x0c4, vacc2x1c4, 2, 6, 3, 7)); in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_arm()
|
D | 4x2c4-minmax-sse.c | 64 __m128 vacc2x0c4 = vacc0x0c4; in xnn_f32_gemm_minmax_ukernel_4x2c4__sse() local 89 vacc2x0c4 = _mm_add_ps(vacc2x0c4, _mm_mul_ps(va2, vb0)); in xnn_f32_gemm_minmax_ukernel_4x2c4__sse() 115 vacc2x0c4 = _mm_add_ps(vacc2x0c4, _mm_mul_ps(_mm_andnot_ps(vmask0, va2), vb0)); in xnn_f32_gemm_minmax_ukernel_4x2c4__sse() 123 …onst __m128 vacc2x01c2 = _mm_add_ps(_mm_unpacklo_ps(vacc2x0c4, vacc2x1c4), _mm_unpackhi_ps(vacc2x0… in xnn_f32_gemm_minmax_ukernel_4x2c4__sse()
|
D | 4x2c4-relu-wasmsimd.c | 64 v128_t vacc2x0c4 = vacc0x0c4; in xnn_f32_gemm_relu_ukernel_4x2c4__wasmsimd() local 89 vacc2x0c4 = wasm_f32x4_add(vacc2x0c4, wasm_f32x4_mul(va2, vb0)); in xnn_f32_gemm_relu_ukernel_4x2c4__wasmsimd() 116 vacc2x0c4 = wasm_f32x4_add(vacc2x0c4, wasm_f32x4_mul(wasm_v128_andnot(va2, vmask0), vb0)); in xnn_f32_gemm_relu_ukernel_4x2c4__wasmsimd() 129 wasm_v32x4_shuffle(vacc2x0c4, vacc2x1c4, 0, 4, 1, 5), in xnn_f32_gemm_relu_ukernel_4x2c4__wasmsimd() 130 wasm_v32x4_shuffle(vacc2x0c4, vacc2x1c4, 2, 6, 3, 7)); in xnn_f32_gemm_relu_ukernel_4x2c4__wasmsimd()
|
D | 4x2c4-minmax-wasmsimd-x86.c | 64 v128_t vacc2x0c4 = vacc0x0c4; in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_x86() local 89 vacc2x0c4 = wasm_f32x4_add(vacc2x0c4, wasm_f32x4_mul(va2, vb0)); in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_x86() 116 vacc2x0c4 = wasm_f32x4_add(vacc2x0c4, wasm_f32x4_mul(wasm_v128_andnot(va2, vmask0), vb0)); in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_x86() 129 wasm_v32x4_shuffle(vacc2x0c4, vacc2x1c4, 0, 4, 1, 5), in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_x86() 130 wasm_v32x4_shuffle(vacc2x0c4, vacc2x1c4, 2, 6, 3, 7)); in xnn_f32_gemm_minmax_ukernel_4x2c4__wasmsimd_x86()
|
/external/XNNPACK/src/f32-igemm/gen/ |
D | 4x2c4-wasmsimd.c | 62 v128_t vacc2x0c4 = vacc0x0c4; in xnn_f32_igemm_ukernel_4x2c4__wasmsimd() local 111 vacc2x0c4 = wasm_f32x4_add(vacc2x0c4, wasm_f32x4_mul(va2, vb0)); in xnn_f32_igemm_ukernel_4x2c4__wasmsimd() 134 vacc2x0c4 = wasm_f32x4_add(vacc2x0c4, wasm_f32x4_mul(wasm_v128_andnot(va2, vmask0), vb0)); in xnn_f32_igemm_ukernel_4x2c4__wasmsimd() 149 wasm_v32x4_shuffle(vacc2x0c4, vacc2x1c4, 0, 4, 1, 5), in xnn_f32_igemm_ukernel_4x2c4__wasmsimd() 150 wasm_v32x4_shuffle(vacc2x0c4, vacc2x1c4, 2, 6, 3, 7)); in xnn_f32_igemm_ukernel_4x2c4__wasmsimd()
|
D | 4x2c4-relu-wasmsimd.c | 62 v128_t vacc2x0c4 = vacc0x0c4; in xnn_f32_igemm_relu_ukernel_4x2c4__wasmsimd() local 111 vacc2x0c4 = wasm_f32x4_add(vacc2x0c4, wasm_f32x4_mul(va2, vb0)); in xnn_f32_igemm_relu_ukernel_4x2c4__wasmsimd() 134 vacc2x0c4 = wasm_f32x4_add(vacc2x0c4, wasm_f32x4_mul(wasm_v128_andnot(va2, vmask0), vb0)); in xnn_f32_igemm_relu_ukernel_4x2c4__wasmsimd() 149 wasm_v32x4_shuffle(vacc2x0c4, vacc2x1c4, 0, 4, 1, 5), in xnn_f32_igemm_relu_ukernel_4x2c4__wasmsimd() 150 wasm_v32x4_shuffle(vacc2x0c4, vacc2x1c4, 2, 6, 3, 7)); in xnn_f32_igemm_relu_ukernel_4x2c4__wasmsimd()
|
D | 4x2c4-minmax-sse.c | 62 __m128 vacc2x0c4 = vacc0x0c4; in xnn_f32_igemm_minmax_ukernel_4x2c4__sse() local 111 vacc2x0c4 = _mm_add_ps(vacc2x0c4, _mm_mul_ps(va2, vb0)); in xnn_f32_igemm_minmax_ukernel_4x2c4__sse() 133 vacc2x0c4 = _mm_add_ps(vacc2x0c4, _mm_mul_ps(_mm_andnot_ps(vmask0, va2), vb0)); in xnn_f32_igemm_minmax_ukernel_4x2c4__sse() 143 …onst __m128 vacc2x01c2 = _mm_add_ps(_mm_unpacklo_ps(vacc2x0c4, vacc2x1c4), _mm_unpackhi_ps(vacc2x0… in xnn_f32_igemm_minmax_ukernel_4x2c4__sse()
|
D | 4x2c4-minmax-wasmsimd-x86.c | 62 v128_t vacc2x0c4 = vacc0x0c4; in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_x86() local 111 vacc2x0c4 = wasm_f32x4_add(vacc2x0c4, wasm_f32x4_mul(va2, vb0)); in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_x86() 134 vacc2x0c4 = wasm_f32x4_add(vacc2x0c4, wasm_f32x4_mul(wasm_v128_andnot(va2, vmask0), vb0)); in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_x86() 149 wasm_v32x4_shuffle(vacc2x0c4, vacc2x1c4, 0, 4, 1, 5), in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_x86() 150 wasm_v32x4_shuffle(vacc2x0c4, vacc2x1c4, 2, 6, 3, 7)); in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_x86()
|
D | 4x2c4-minmax-wasmsimd-arm.c | 64 v128_t vacc2x0c4 = vacc0x0c4; in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_arm() local 113 vacc2x0c4 = wasm_f32x4_add(vacc2x0c4, wasm_f32x4_mul(va2, vb0)); in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_arm() 136 vacc2x0c4 = wasm_f32x4_add(vacc2x0c4, wasm_f32x4_mul(wasm_v128_andnot(va2, vmask0), vb0)); in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_arm() 151 wasm_v32x4_shuffle(vacc2x0c4, vacc2x1c4, 0, 4, 1, 5), in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_arm() 152 wasm_v32x4_shuffle(vacc2x0c4, vacc2x1c4, 2, 6, 3, 7)); in xnn_f32_igemm_minmax_ukernel_4x2c4__wasmsimd_arm()
|