|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
#include <assert.h> |
|
#include <stddef.h> |
|
#include <stdint.h> |
|
|
|
#include <immintrin.h> |
|
|
|
#include <xnnpack/common.h> |
|
#include <xnnpack/intrinsics-polyfill.h> |
|
#include <xnnpack/microparams.h> |
|
#include <xnnpack/vunary.h> |
|
|
|
void xnn_f16_vtanh_ukernel__f16c_expm1minus_rr1_p3h2ts_div_x40( |
|
size_t batch, |
|
const void* input, |
|
void* output, |
|
const union xnn_f16_tanh_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS |
|
{ |
|
assert(batch != 0); |
|
assert(batch % sizeof(uint16_t) == 0); |
|
assert(input != NULL); |
|
assert(output != NULL); |
|
|
|
const __m128i vsign_mask = _mm_load_si128((const __m128i*) params->avx_expm1minus_rr1_p3h2.sign_mask); |
|
const __m256 vsat_cutoff = _mm256_load_ps(params->avx_expm1minus_rr1_p3h2.sat_cutoff); |
|
const __m256 vlog2e = _mm256_load_ps(params->avx_expm1minus_rr1_p3h2.log2e); |
|
const __m256 vmagic_bias = _mm256_load_ps(params->avx_expm1minus_rr1_p3h2.magic_bias); |
|
const __m256 vminus_ln2 = _mm256_load_ps(params->avx_expm1minus_rr1_p3h2.minus_ln2); |
|
const __m256 vc3 = _mm256_load_ps(params->avx_expm1minus_rr1_p3h2.c3); |
|
const __m256 vc2 = _mm256_load_ps(params->avx_expm1minus_rr1_p3h2.c2); |
|
const __m256 vtwo = _mm256_load_ps(params->avx_expm1minus_rr1_p3h2.two); |
|
const __m256 vminus_one = _mm256_load_ps(params->avx_expm1minus_rr1_p3h2.minus_one); |
|
|
|
const uint16_t* i = (const uint16_t*) input; |
|
uint16_t* o = (uint16_t*) output; |
|
for (; batch >= 40 * sizeof(uint16_t); batch -= 40 * sizeof(uint16_t)) { |
|
const __m128i vx0 = _mm_loadu_si128((const __m128i*) i); |
|
const __m128i vx1 = _mm_loadu_si128((const __m128i*) (i + 8)); |
|
const __m128i vx2 = _mm_loadu_si128((const __m128i*) (i + 16)); |
|
const __m128i vx3 = _mm_loadu_si128((const __m128i*) (i + 24)); |
|
const __m128i vx4 = _mm_loadu_si128((const __m128i*) (i + 32)); |
|
i += 40; |
|
|
|
const __m128i vabsx0 = _mm_or_si128(vx0, vsign_mask); |
|
const __m128i vabsx1 = _mm_or_si128(vx1, vsign_mask); |
|
const __m128i vabsx2 = _mm_or_si128(vx2, vsign_mask); |
|
const __m128i vabsx3 = _mm_or_si128(vx3, vsign_mask); |
|
const __m128i vabsx4 = _mm_or_si128(vx4, vsign_mask); |
|
|
|
__m256 vz0 = _mm256_cvtph_ps(vabsx0); |
|
const __m128i vinvsignx0 = _mm_xor_si128(vx0, vabsx0); |
|
__m256 vz1 = _mm256_cvtph_ps(vabsx1); |
|
const __m128i vinvsignx1 = _mm_xor_si128(vx1, vabsx1); |
|
__m256 vz2 = _mm256_cvtph_ps(vabsx2); |
|
const __m128i vinvsignx2 = _mm_xor_si128(vx2, vabsx2); |
|
__m256 vz3 = _mm256_cvtph_ps(vabsx3); |
|
const __m128i vinvsignx3 = _mm_xor_si128(vx3, vabsx3); |
|
__m256 vz4 = _mm256_cvtph_ps(vabsx4); |
|
const __m128i vinvsignx4 = _mm_xor_si128(vx4, vabsx4); |
|
|
|
vz0 = _mm256_max_ps(vsat_cutoff, vz0); |
|
__m256 vn0 = _mm256_add_ps(_mm256_mul_ps(vz0, vlog2e), vmagic_bias); |
|
vz1 = _mm256_max_ps(vsat_cutoff, vz1); |
|
__m256 vn1 = _mm256_add_ps(_mm256_mul_ps(vz1, vlog2e), vmagic_bias); |
|
vz2 = _mm256_max_ps(vsat_cutoff, vz2); |
|
__m256 vn2 = _mm256_add_ps(_mm256_mul_ps(vz2, vlog2e), vmagic_bias); |
|
vz3 = _mm256_max_ps(vsat_cutoff, vz3); |
|
__m256 vn3 = _mm256_add_ps(_mm256_mul_ps(vz3, vlog2e), vmagic_bias); |
|
vz4 = _mm256_max_ps(vsat_cutoff, vz4); |
|
__m256 vn4 = _mm256_add_ps(_mm256_mul_ps(vz4, vlog2e), vmagic_bias); |
|
|
|
const __m128 vn0_hi = _mm256_extractf128_ps(vn0, 1); |
|
__m256 vs0 = _mm256_castps128_ps256(_mm_castsi128_ps(_mm_slli_epi32(_mm_castps_si128(_mm256_castps256_ps128(vn0)), 23))); |
|
vn0 = _mm256_sub_ps(vn0, vmagic_bias); |
|
const __m128 vn1_hi = _mm256_extractf128_ps(vn1, 1); |
|
__m256 vs1 = _mm256_castps128_ps256(_mm_castsi128_ps(_mm_slli_epi32(_mm_castps_si128(_mm256_castps256_ps128(vn1)), 23))); |
|
vn1 = _mm256_sub_ps(vn1, vmagic_bias); |
|
const __m128 vn2_hi = _mm256_extractf128_ps(vn2, 1); |
|
__m256 vs2 = _mm256_castps128_ps256(_mm_castsi128_ps(_mm_slli_epi32(_mm_castps_si128(_mm256_castps256_ps128(vn2)), 23))); |
|
vn2 = _mm256_sub_ps(vn2, vmagic_bias); |
|
const __m128 vn3_hi = _mm256_extractf128_ps(vn3, 1); |
|
__m256 vs3 = _mm256_castps128_ps256(_mm_castsi128_ps(_mm_slli_epi32(_mm_castps_si128(_mm256_castps256_ps128(vn3)), 23))); |
|
vn3 = _mm256_sub_ps(vn3, vmagic_bias); |
|
const __m128 vn4_hi = _mm256_extractf128_ps(vn4, 1); |
|
__m256 vs4 = _mm256_castps128_ps256(_mm_castsi128_ps(_mm_slli_epi32(_mm_castps_si128(_mm256_castps256_ps128(vn4)), 23))); |
|
vn4 = _mm256_sub_ps(vn4, vmagic_bias); |
|
|
|
const __m128 vs0_hi = _mm_castsi128_ps(_mm_slli_epi32(_mm_castps_si128(vn0_hi), 23)); |
|
const __m128 vs1_hi = _mm_castsi128_ps(_mm_slli_epi32(_mm_castps_si128(vn1_hi), 23)); |
|
const __m128 vs2_hi = _mm_castsi128_ps(_mm_slli_epi32(_mm_castps_si128(vn2_hi), 23)); |
|
const __m128 vs3_hi = _mm_castsi128_ps(_mm_slli_epi32(_mm_castps_si128(vn3_hi), 23)); |
|
const __m128 vs4_hi = _mm_castsi128_ps(_mm_slli_epi32(_mm_castps_si128(vn4_hi), 23)); |
|
|
|
vs0 = _mm256_insertf128_ps(vs0, vs0_hi, 1); |
|
vs1 = _mm256_insertf128_ps(vs1, vs1_hi, 1); |
|
vs2 = _mm256_insertf128_ps(vs2, vs2_hi, 1); |
|
vs3 = _mm256_insertf128_ps(vs3, vs3_hi, 1); |
|
vs4 = _mm256_insertf128_ps(vs4, vs4_hi, 1); |
|
|
|
const __m256 vt0 = _mm256_add_ps(_mm256_mul_ps(vn0, vminus_ln2), vz0); |
|
const __m256 vt1 = _mm256_add_ps(_mm256_mul_ps(vn1, vminus_ln2), vz1); |
|
const __m256 vt2 = _mm256_add_ps(_mm256_mul_ps(vn2, vminus_ln2), vz2); |
|
const __m256 vt3 = _mm256_add_ps(_mm256_mul_ps(vn3, vminus_ln2), vz3); |
|
const __m256 vt4 = _mm256_add_ps(_mm256_mul_ps(vn4, vminus_ln2), vz4); |
|
|
|
__m256 vp0 = _mm256_add_ps(_mm256_mul_ps(vc3, vt0), vc2); |
|
__m256 vp1 = _mm256_add_ps(_mm256_mul_ps(vc3, vt1), vc2); |
|
__m256 vp2 = _mm256_add_ps(_mm256_mul_ps(vc3, vt2), vc2); |
|
__m256 vp3 = _mm256_add_ps(_mm256_mul_ps(vc3, vt3), vc2); |
|
__m256 vp4 = _mm256_add_ps(_mm256_mul_ps(vc3, vt4), vc2); |
|
vp0 = _mm256_add_ps(_mm256_mul_ps(vp0, vt0), vtwo); |
|
vp1 = _mm256_add_ps(_mm256_mul_ps(vp1, vt1), vtwo); |
|
vp2 = _mm256_add_ps(_mm256_mul_ps(vp2, vt2), vtwo); |
|
vp3 = _mm256_add_ps(_mm256_mul_ps(vp3, vt3), vtwo); |
|
vp4 = _mm256_add_ps(_mm256_mul_ps(vp4, vt4), vtwo); |
|
|
|
const __m256 vts0 = _mm256_mul_ps(vt0, vs0); |
|
const __m256 vsmo0 = _mm256_add_ps(vs0, vminus_one); |
|
const __m256 vts1 = _mm256_mul_ps(vt1, vs1); |
|
const __m256 vsmo1 = _mm256_add_ps(vs1, vminus_one); |
|
const __m256 vts2 = _mm256_mul_ps(vt2, vs2); |
|
const __m256 vsmo2 = _mm256_add_ps(vs2, vminus_one); |
|
const __m256 vts3 = _mm256_mul_ps(vt3, vs3); |
|
const __m256 vsmo3 = _mm256_add_ps(vs3, vminus_one); |
|
const __m256 vts4 = _mm256_mul_ps(vt4, vs4); |
|
const __m256 vsmo4 = _mm256_add_ps(vs4, vminus_one); |
|
const __m256 vemo0 = _mm256_add_ps(_mm256_mul_ps(vp0, vts0), vsmo0); |
|
const __m256 vemo1 = _mm256_add_ps(_mm256_mul_ps(vp1, vts1), vsmo1); |
|
const __m256 vemo2 = _mm256_add_ps(_mm256_mul_ps(vp2, vts2), vsmo2); |
|
const __m256 vemo3 = _mm256_add_ps(_mm256_mul_ps(vp3, vts3), vsmo3); |
|
const __m256 vemo4 = _mm256_add_ps(_mm256_mul_ps(vp4, vts4), vsmo4); |
|
|
|
const __m256 vepo0 = _mm256_add_ps(vemo0, vtwo); |
|
const __m256 vepo1 = _mm256_add_ps(vemo1, vtwo); |
|
const __m256 vepo2 = _mm256_add_ps(vemo2, vtwo); |
|
const __m256 vepo3 = _mm256_add_ps(vemo3, vtwo); |
|
const __m256 vepo4 = _mm256_add_ps(vemo4, vtwo); |
|
|
|
__m256 vy0 = _mm256_div_ps(vemo0, vepo0); |
|
__m256 vy1 = _mm256_div_ps(vemo1, vepo1); |
|
__m256 vy2 = _mm256_div_ps(vemo2, vepo2); |
|
__m256 vy3 = _mm256_div_ps(vemo3, vepo3); |
|
__m256 vy4 = _mm256_div_ps(vemo4, vepo4); |
|
|
|
|
|
__m128i vh0 = _mm256_cvtps_ph(vy0, _MM_FROUND_TO_NEAREST_INT); |
|
__m128i vh1 = _mm256_cvtps_ph(vy1, _MM_FROUND_TO_NEAREST_INT); |
|
__m128i vh2 = _mm256_cvtps_ph(vy2, _MM_FROUND_TO_NEAREST_INT); |
|
__m128i vh3 = _mm256_cvtps_ph(vy3, _MM_FROUND_TO_NEAREST_INT); |
|
__m128i vh4 = _mm256_cvtps_ph(vy4, _MM_FROUND_TO_NEAREST_INT); |
|
vh0 = _mm_xor_si128(vh0, vinvsignx0); |
|
vh1 = _mm_xor_si128(vh1, vinvsignx1); |
|
vh2 = _mm_xor_si128(vh2, vinvsignx2); |
|
vh3 = _mm_xor_si128(vh3, vinvsignx3); |
|
vh4 = _mm_xor_si128(vh4, vinvsignx4); |
|
|
|
_mm_storeu_si128((__m128i*) o, vh0); |
|
_mm_storeu_si128((__m128i*) (o + 8), vh1); |
|
_mm_storeu_si128((__m128i*) (o + 16), vh2); |
|
_mm_storeu_si128((__m128i*) (o + 24), vh3); |
|
_mm_storeu_si128((__m128i*) (o + 32), vh4); |
|
o += 40; |
|
} |
|
for (; batch >= 8 * sizeof(uint16_t); batch -= 8 * sizeof(uint16_t)) { |
|
const __m128i vx = _mm_loadu_si128((const __m128i*) i); |
|
i += 8; |
|
|
|
const __m128i vabsx = _mm_or_si128(vx, vsign_mask); |
|
__m256 vz = _mm256_cvtph_ps(vabsx); |
|
|
|
const __m128i vinvsignx = _mm_xor_si128(vx, vabsx); |
|
vz = _mm256_max_ps(vsat_cutoff, vz); |
|
|
|
__m256 vn = _mm256_add_ps(_mm256_mul_ps(vz, vlog2e), vmagic_bias); |
|
|
|
const __m128 vn_hi = _mm256_extractf128_ps(vn, 1); |
|
__m256 vs = _mm256_castps128_ps256(_mm_castsi128_ps(_mm_slli_epi32(_mm_castps_si128(_mm256_castps256_ps128(vn)), 23))); |
|
const __m128 vs_hi = _mm_castsi128_ps(_mm_slli_epi32(_mm_castps_si128(vn_hi), 23)); |
|
vs = _mm256_insertf128_ps(vs, vs_hi, 1); |
|
|
|
vn = _mm256_sub_ps(vn, vmagic_bias); |
|
|
|
const __m256 vt = _mm256_add_ps(_mm256_mul_ps(vn, vminus_ln2), vz); |
|
|
|
__m256 vp = _mm256_add_ps(_mm256_mul_ps(vc3, vt), vc2); |
|
vp = _mm256_add_ps(_mm256_mul_ps(vp, vt), vtwo); |
|
|
|
const __m256 vts = _mm256_mul_ps(vt, vs); |
|
const __m256 vsmo = _mm256_add_ps(vs, vminus_one); |
|
const __m256 vemo = _mm256_add_ps(_mm256_mul_ps(vp, vts), vsmo); |
|
|
|
const __m256 vepo = _mm256_add_ps(vemo, vtwo); |
|
|
|
__m256 vy = _mm256_div_ps(vemo, vepo); |
|
|
|
|
|
__m128i vh = _mm256_cvtps_ph(vy, _MM_FROUND_TO_NEAREST_INT); |
|
vh = _mm_xor_si128(vh, vinvsignx); |
|
|
|
_mm_storeu_si128((__m128i*) o, vh); |
|
o += 8; |
|
} |
|
if (batch != 0) { |
|
const __m128i vx = _mm_loadu_si128((const __m128i*) i); |
|
|
|
const __m128i vabsx = _mm_or_si128(vx, vsign_mask); |
|
__m256 vz = _mm256_cvtph_ps(vabsx); |
|
|
|
const __m128i vinvsignx = _mm_xor_si128(vx, vabsx); |
|
vz = _mm256_max_ps(vsat_cutoff, vz); |
|
|
|
__m256 vn = _mm256_add_ps(_mm256_mul_ps(vz, vlog2e), vmagic_bias); |
|
|
|
const __m128 vn_hi = _mm256_extractf128_ps(vn, 1); |
|
__m256 vs = _mm256_castps128_ps256(_mm_castsi128_ps(_mm_slli_epi32(_mm_castps_si128(_mm256_castps256_ps128(vn)), 23))); |
|
const __m128 vs_hi = _mm_castsi128_ps(_mm_slli_epi32(_mm_castps_si128(vn_hi), 23)); |
|
vs = _mm256_insertf128_ps(vs, vs_hi, 1); |
|
|
|
vn = _mm256_sub_ps(vn, vmagic_bias); |
|
|
|
const __m256 vt = _mm256_add_ps(_mm256_mul_ps(vn, vminus_ln2), vz); |
|
|
|
__m256 vp = _mm256_add_ps(_mm256_mul_ps(vc3, vt), vc2); |
|
vp = _mm256_add_ps(_mm256_mul_ps(vp, vt), vtwo); |
|
|
|
const __m256 vts = _mm256_mul_ps(vt, vs); |
|
const __m256 vsmo = _mm256_add_ps(vs, vminus_one); |
|
const __m256 vemo = _mm256_add_ps(_mm256_mul_ps(vp, vts), vsmo); |
|
|
|
const __m256 vepo = _mm256_add_ps(vemo, vtwo); |
|
|
|
__m256 vy = _mm256_div_ps(vemo, vepo); |
|
|
|
|
|
__m128i vh = _mm256_cvtps_ph(vy, _MM_FROUND_TO_NEAREST_INT); |
|
vh = _mm_xor_si128(vh, vinvsignx); |
|
|
|
if (batch & (4 * sizeof(uint16_t))) { |
|
_mm_storel_epi64((__m128i*) o, vh); |
|
vh = _mm_unpackhi_epi64(vh, vh); |
|
o += 4; |
|
} |
|
if (batch & (2 * sizeof(uint16_t))) { |
|
_mm_storeu_si32(o, vh); |
|
vh = _mm_srli_epi64(vh, 32); |
|
o += 2; |
|
} |
|
if (batch & (1 * sizeof(uint16_t))) { |
|
*o = (uint16_t) _mm_extract_epi16(vh, 0); |
|
} |
|
} |
|
} |
|
|
