|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
#include <assert.h> |
|
|
|
#include <arm_neon.h> |
|
|
|
#include <xnnpack/gemm.h> |
|
|
|
|
|
void xnn_f32_gemm_minmax_ukernel_1x16__aarch64_neonfma_lane_ld128( |
|
size_t mr, |
|
size_t nc, |
|
size_t kc, |
|
const float* restrict a, |
|
size_t a_stride, |
|
const float* restrict w, |
|
float* restrict c, |
|
size_t cm_stride, |
|
size_t cn_stride, |
|
const union xnn_f32_minmax_params params[restrict XNN_MIN_ELEMENTS(1)]) |
|
{ |
|
assert(mr != 0); |
|
assert(mr <= 1); |
|
assert(nc != 0); |
|
assert(kc != 0); |
|
assert(kc % sizeof(float) == 0); |
|
assert(a != NULL); |
|
assert(w != NULL); |
|
assert(c != NULL); |
|
|
|
const float* a0 = a; |
|
float* c0 = c; |
|
|
|
do { |
|
float32x4_t vacc0x0123 = vld1q_f32(w); w += 4; |
|
float32x4_t vacc0x4567 = vld1q_f32(w); w += 4; |
|
float32x4_t vacc0x89AB = vld1q_f32(w); w += 4; |
|
float32x4_t vacc0xCDEF = vld1q_f32(w); w += 4; |
|
|
|
size_t k = kc; |
|
if XNN_LIKELY(k >= 4 * sizeof(float)) { |
|
do { |
|
const float32x4_t va0 = vld1q_f32(a0); a0 += 4; |
|
|
|
|
|
const float32x4_t vb0123c0 = vld1q_f32(w); w += 4; |
|
const float32x4_t vb4567c0 = vld1q_f32(w); w += 4; |
|
const float32x4_t vb89ABc0 = vld1q_f32(w); w += 4; |
|
const float32x4_t vbCDEFc0 = vld1q_f32(w); w += 4; |
|
|
|
vacc0x0123 = vfmaq_lane_f32(vacc0x0123, vb0123c0, vget_low_f32(va0), 0); |
|
vacc0x4567 = vfmaq_lane_f32(vacc0x4567, vb4567c0, vget_low_f32(va0), 0); |
|
vacc0x89AB = vfmaq_lane_f32(vacc0x89AB, vb89ABc0, vget_low_f32(va0), 0); |
|
vacc0xCDEF = vfmaq_lane_f32(vacc0xCDEF, vbCDEFc0, vget_low_f32(va0), 0); |
|
|
|
const float32x4_t vb0123c1 = vld1q_f32(w); w += 4; |
|
const float32x4_t vb4567c1 = vld1q_f32(w); w += 4; |
|
const float32x4_t vb89ABc1 = vld1q_f32(w); w += 4; |
|
const float32x4_t vbCDEFc1 = vld1q_f32(w); w += 4; |
|
|
|
vacc0x0123 = vfmaq_lane_f32(vacc0x0123, vb0123c1, vget_low_f32(va0), 1); |
|
vacc0x4567 = vfmaq_lane_f32(vacc0x4567, vb4567c1, vget_low_f32(va0), 1); |
|
vacc0x89AB = vfmaq_lane_f32(vacc0x89AB, vb89ABc1, vget_low_f32(va0), 1); |
|
vacc0xCDEF = vfmaq_lane_f32(vacc0xCDEF, vbCDEFc1, vget_low_f32(va0), 1); |
|
|
|
const float32x4_t vb0123c2 = vld1q_f32(w); w += 4; |
|
const float32x4_t vb4567c2 = vld1q_f32(w); w += 4; |
|
const float32x4_t vb89ABc2 = vld1q_f32(w); w += 4; |
|
const float32x4_t vbCDEFc2 = vld1q_f32(w); w += 4; |
|
|
|
vacc0x0123 = vfmaq_lane_f32(vacc0x0123, vb0123c2, vget_high_f32(va0), 0); |
|
vacc0x4567 = vfmaq_lane_f32(vacc0x4567, vb4567c2, vget_high_f32(va0), 0); |
|
vacc0x89AB = vfmaq_lane_f32(vacc0x89AB, vb89ABc2, vget_high_f32(va0), 0); |
|
vacc0xCDEF = vfmaq_lane_f32(vacc0xCDEF, vbCDEFc2, vget_high_f32(va0), 0); |
|
|
|
const float32x4_t vb0123c3 = vld1q_f32(w); w += 4; |
|
const float32x4_t vb4567c3 = vld1q_f32(w); w += 4; |
|
const float32x4_t vb89ABc3 = vld1q_f32(w); w += 4; |
|
const float32x4_t vbCDEFc3 = vld1q_f32(w); w += 4; |
|
|
|
vacc0x0123 = vfmaq_lane_f32(vacc0x0123, vb0123c3, vget_high_f32(va0), 1); |
|
vacc0x4567 = vfmaq_lane_f32(vacc0x4567, vb4567c3, vget_high_f32(va0), 1); |
|
vacc0x89AB = vfmaq_lane_f32(vacc0x89AB, vb89ABc3, vget_high_f32(va0), 1); |
|
vacc0xCDEF = vfmaq_lane_f32(vacc0xCDEF, vbCDEFc3, vget_high_f32(va0), 1); |
|
k -= 4 * sizeof(float); |
|
} while (k >= 4 * sizeof(float)); |
|
} |
|
|
|
if XNN_UNLIKELY(k != 0) { |
|
if XNN_UNLIKELY(k & (2 * sizeof(float))) { |
|
const float32x2_t va0 = vld1_f32(a0); a0 += 2; |
|
|
|
|
|
const float32x4_t vb0123c0 = vld1q_f32(w); w += 4; |
|
const float32x4_t vb4567c0 = vld1q_f32(w); w += 4; |
|
const float32x4_t vb89ABc0 = vld1q_f32(w); w += 4; |
|
const float32x4_t vbCDEFc0 = vld1q_f32(w); w += 4; |
|
|
|
vacc0x0123 = vfmaq_lane_f32(vacc0x0123, vb0123c0, va0, 0); |
|
vacc0x4567 = vfmaq_lane_f32(vacc0x4567, vb4567c0, va0, 0); |
|
vacc0x89AB = vfmaq_lane_f32(vacc0x89AB, vb89ABc0, va0, 0); |
|
vacc0xCDEF = vfmaq_lane_f32(vacc0xCDEF, vbCDEFc0, va0, 0); |
|
|
|
const float32x4_t vb0123c1 = vld1q_f32(w); w += 4; |
|
const float32x4_t vb4567c1 = vld1q_f32(w); w += 4; |
|
const float32x4_t vb89ABc1 = vld1q_f32(w); w += 4; |
|
const float32x4_t vbCDEFc1 = vld1q_f32(w); w += 4; |
|
|
|
vacc0x0123 = vfmaq_lane_f32(vacc0x0123, vb0123c1, va0, 1); |
|
vacc0x4567 = vfmaq_lane_f32(vacc0x4567, vb4567c1, va0, 1); |
|
vacc0x89AB = vfmaq_lane_f32(vacc0x89AB, vb89ABc1, va0, 1); |
|
vacc0xCDEF = vfmaq_lane_f32(vacc0xCDEF, vbCDEFc1, va0, 1); |
|
} |
|
if XNN_UNLIKELY(k & (1 * sizeof(float))) { |
|
const float32x4_t va0 = vld1q_dup_f32(a0); a0 += 1; |
|
|
|
const float32x4_t vb0123 = vld1q_f32(w); w += 4; |
|
const float32x4_t vb4567 = vld1q_f32(w); w += 4; |
|
const float32x4_t vb89AB = vld1q_f32(w); w += 4; |
|
const float32x4_t vbCDEF = vld1q_f32(w); w += 4; |
|
|
|
vacc0x0123 = vfmaq_f32(vacc0x0123, va0, vb0123); |
|
vacc0x4567 = vfmaq_f32(vacc0x4567, va0, vb4567); |
|
vacc0x89AB = vfmaq_f32(vacc0x89AB, va0, vb89AB); |
|
vacc0xCDEF = vfmaq_f32(vacc0xCDEF, va0, vbCDEF); |
|
} |
|
} |
|
const float32x4_t vmax = vld1q_dup_f32(¶ms->scalar.max); |
|
vacc0x0123 = vminq_f32(vacc0x0123, vmax); |
|
vacc0x4567 = vminq_f32(vacc0x4567, vmax); |
|
vacc0x89AB = vminq_f32(vacc0x89AB, vmax); |
|
vacc0xCDEF = vminq_f32(vacc0xCDEF, vmax); |
|
|
|
const float32x4_t vmin = vld1q_dup_f32(¶ms->scalar.min); |
|
vacc0x0123 = vmaxq_f32(vacc0x0123, vmin); |
|
vacc0x4567 = vmaxq_f32(vacc0x4567, vmin); |
|
vacc0x89AB = vmaxq_f32(vacc0x89AB, vmin); |
|
vacc0xCDEF = vmaxq_f32(vacc0xCDEF, vmin); |
|
|
|
if XNN_LIKELY(nc >= 16) { |
|
vst1q_f32(c0, vacc0x0123); |
|
vst1q_f32(c0 + 4, vacc0x4567); |
|
vst1q_f32(c0 + 8, vacc0x89AB); |
|
vst1q_f32(c0 + 12, vacc0xCDEF); |
|
c0 = (float*) ((uintptr_t) c0 + cn_stride); |
|
|
|
a0 = (const float*) ((uintptr_t) a0 - kc); |
|
|
|
nc -= 16; |
|
|
|
} else { |
|
if (nc & 8) { |
|
vst1q_f32(c0, vacc0x0123); c0 += 4; |
|
vst1q_f32(c0, vacc0x4567); c0 += 4; |
|
|
|
vacc0x0123 = vacc0x89AB; |
|
vacc0x4567 = vacc0xCDEF; |
|
} |
|
if (nc & 4) { |
|
vst1q_f32(c0, vacc0x0123); c0 += 4; |
|
|
|
vacc0x0123 = vacc0x4567; |
|
vacc0x4567 = vacc0x89AB; |
|
vacc0x89AB = vacc0xCDEF; |
|
} |
|
float32x2_t vacc0x01 = vget_low_f32(vacc0x0123); |
|
if (nc & 2) { |
|
vst1_f32(c0, vacc0x01); c0 += 2; |
|
|
|
vacc0x01 = vget_high_f32(vacc0x0123); |
|
} |
|
if (nc & 1) { |
|
vst1_lane_f32(c0, vacc0x01, 0); |
|
} |
|
|
|
nc = 0; |
|
} |
|
} while (nc != 0); |
|
} |
|
|