| // Auto-generated file. Do not edit! |
| // Template: src/f32-vlrelu/wasmsimd-minmax.c.in |
| // Generator: tools/xngen |
| // |
| // Copyright 2020 Google LLC |
| // |
| // This source code is licensed under the BSD-style license found in the |
| // LICENSE file in the root directory of this source tree. |
| |
| #include <assert.h> |
| |
| #include <wasm_simd128.h> |
| |
| #include <xnnpack/common.h> |
| #include <xnnpack/vunary.h> |
| |
| |
| void xnn_f32_vlrelu_ukernel__wasmsimd_minmax_x4( |
| size_t n, |
| const float* x, |
| float* y, |
| const union xnn_f32_lrelu_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_DISABLE_TSAN |
| { |
| assert(n != 0); |
| assert(n % sizeof(float) == 0); |
| |
| const v128_t vslope = wasm_v32x4_load_splat(¶ms->scalar.slope); |
| const v128_t vzero = wasm_f32x4_splat(0.0f); |
| for (; n >= 4 * sizeof(float); n -= 4 * sizeof(float)) { |
| v128_t vx = wasm_v128_load(x); |
| x += 4; |
| v128_t vacc = wasm_i32x4_max(vx, vzero); |
| vx = wasm_i32x4_min(vx, vzero); |
| vacc = wasm_f32x4_add(vacc, wasm_f32x4_mul(vx, vslope)); |
| wasm_v128_store(y, vacc); |
| y += 4; |
| } |
| if XNN_UNLIKELY(n != 0) { |
| v128_t vx = wasm_v128_load(x); |
| v128_t vacc = wasm_i32x4_max(vx, vzero); |
| vx = wasm_i32x4_min(vx, vzero); |
| vacc = wasm_f32x4_add(vacc, wasm_f32x4_mul(vx, vslope)); |
| |
| if (n & (2 * sizeof(float))) { |
| *((double*) y) = wasm_f64x2_extract_lane(vacc, 0); |
| vacc = wasm_v32x4_shuffle(vacc, vacc, 2, 3, 2, 3); |
| y += 2; |
| } |
| if (n & (1 * sizeof(float))) { |
| *y = wasm_f32x4_extract_lane(vacc, 0); |
| } |
| } |
| } |