1 // Auto-generated file. Do not edit!
2 // Template: src/f32-vscaleextexp/avx512f-p5-scalef.c.in
3 // Generator: tools/xngen
4 //
5 // Copyright 2019 Google LLC
6 //
7 // This source code is licensed under the BSD-style license found in the
8 // LICENSE file in the root directory of this source tree.
9
10 #include <assert.h>
11
12 #include <immintrin.h>
13
14 #include <xnnpack/common.h>
15 #include <xnnpack/intrinsics-polyfill.h>
16 #include <xnnpack/vscaleextexp.h>
17
18
xnn_f32_vscaleextexp_ukernel__avx512f_p5_scalef_x32(size_t elements,const float * x,float * y,float scale_value,float scale_exp)19 void xnn_f32_vscaleextexp_ukernel__avx512f_p5_scalef_x32(
20 size_t elements,
21 const float* x,
22 float* y,
23 float scale_value,
24 float scale_exp)
25 {
26 assert(elements % sizeof(float) == 0);
27
28 const __m512 vlog2e = _mm512_set1_ps(0x1.715476p+0f);
29 const __m512 vminus_ln2_hi = _mm512_set1_ps(-0x1.62E43p-1f);
30 const __m512 vminus_ln2_lo = _mm512_set1_ps(0x1.05C61p-29f);
31
32 const __m512 vc0 = _mm512_set1_ps(1.0f);
33 const __m512 vc1 = _mm512_set1_ps(0x1.FFFFF6p-1f);
34 const __m512 vc2 = _mm512_set1_ps(0x1.FFFDC6p-2f);
35 const __m512 vc3 = _mm512_set1_ps(0x1.555A80p-3f);
36 const __m512 vc4 = _mm512_set1_ps(0x1.573A1Ap-5f);
37 const __m512 vc5 = _mm512_set1_ps(0x1.0F9F9Cp-7f);
38
39 const __m512 vscalev = _mm512_set1_ps(scale_value);
40 const __m512 vscalee = _mm512_set1_ps(scale_exp);
41
42 for (; elements >= 32 * sizeof(float); elements -= 32 * sizeof(float)) {
43 // Load 32 (2x16) inputs at a time.
44 const __m512 vx0 = _mm512_loadu_ps(x);
45 const __m512 vx1 = _mm512_loadu_ps(x + 16);
46 x += 32;
47
48 // Compute reduced argument elements := round(x / log(2)).
49 const __m512 vn0 = _mm512_roundscale_ps(_mm512_mul_ps(vx0, vlog2e), 0);
50 const __m512 vn1 = _mm512_roundscale_ps(_mm512_mul_ps(vx1, vlog2e), 0);
51
52 // Compute reduced argument t := x - elements * log(2).
53 // Use Cody-Waite range reduction method (note two constants to represent log(2)) to improve accuracy.
54 __m512 vt0 = _mm512_fmadd_ps(vn0, vminus_ln2_hi, vx0);
55 __m512 vt1 = _mm512_fmadd_ps(vn1, vminus_ln2_hi, vx1);
56
57 vt0 = _mm512_fmadd_ps(vn0, vminus_ln2_lo, vt0);
58 vt1 = _mm512_fmadd_ps(vn1, vminus_ln2_lo, vt1);
59
60 // Compute degree-5 polynomial approximation for exp(t) on [-log(2)/2, log(2)/2].
61 __m512 vp0 = _mm512_fmadd_ps(vc5, vt0, vc4);
62 __m512 vp1 = _mm512_fmadd_ps(vc5, vt1, vc4);
63
64 vp0 = _mm512_fmadd_ps(vp0, vt0, vc3);
65 vp1 = _mm512_fmadd_ps(vp1, vt1, vc3);
66
67 vp0 = _mm512_fmadd_ps(vp0, vt0, vc2);
68 vp1 = _mm512_fmadd_ps(vp1, vt1, vc2);
69
70 vp0 = _mm512_fmadd_ps(vp0, vt0, vc1);
71 vp1 = _mm512_fmadd_ps(vp1, vt1, vc1);
72
73 vp0 = _mm512_fmadd_ps(vp0, vt0, vc0);
74 vp1 = _mm512_fmadd_ps(vp1, vt1, vc0);
75
76 // Multiply "extended" floating-point numbers in ("mantissa", "exponent") representation where
77 // - vnX is "exponent"
78 // - vpX is "mantissa"
79 //
80 // exp2(ae) * av * exp2(be) * bv =
81 // = exp2(ae + be) * (av * bv)
82 __m512 vf0 = _mm512_mul_ps(vp0, vscalev);
83 __m512 vf1 = _mm512_mul_ps(vp1, vscalev);
84
85 const __m512 ve0 = _mm512_add_ps(vn0, vscalee);
86 const __m512 ve1 = _mm512_add_ps(vn1, vscalee);
87
88 // Multiply "mantissa" by the exp2("exponent").
89 vf0 = _mm512_scalef_ps(vf0, ve0);
90 vf1 = _mm512_scalef_ps(vf1, ve1);
91
92 // Store 128 (8x16) results at a time.
93 _mm512_storeu_ps(y, vf0);
94 _mm512_storeu_ps(y + 0, vf0);
95 _mm512_storeu_ps(y + 16, vf1);
96 y += 32;
97 }
98
99 for (; elements >= 16 * sizeof(float); elements -= 16 * sizeof(float)) {
100 // Load 16 inputs at a time.
101 const __m512 vx = _mm512_loadu_ps(x);
102 x += 16;
103
104 // Compute reduced argument elements := round(x / log(2)).
105 const __m512 vn = _mm512_roundscale_ps(_mm512_mul_ps(vx, vlog2e), 0);
106
107 // Compute reduced argument t := x - elements * log(2).
108 // Use Cody-Waite range reduction method (note two constants to represent log(2)) to improve accuracy.
109 __m512 vt = _mm512_fmadd_ps(vn, vminus_ln2_hi, vx);
110 vt = _mm512_fmadd_ps(vn, vminus_ln2_lo, vt);
111
112 // Compute degree-5 polynomial approximation for exp(t) on [-log(2)/2, log(2)/2].
113 __m512 vp = _mm512_fmadd_ps(vc5, vt, vc4);
114 vp = _mm512_fmadd_ps(vp, vt, vc3);
115 vp = _mm512_fmadd_ps(vp, vt, vc2);
116 vp = _mm512_fmadd_ps(vp, vt, vc1);
117 vp = _mm512_fmadd_ps(vp, vt, vc0);
118
119 // Multiply "extended" floating-point numbers in ("mantissa", "exponent") representation.
120 __m512 vf = _mm512_mul_ps(vp, vscalev);
121 const __m512 ve = _mm512_add_ps(vn, vscalee);
122
123 // Multiply "mantissa" by the exp2("exponent").
124 vf = _mm512_scalef_ps(vf, ve);
125
126 // Store 16 results at a time.
127 _mm512_storeu_ps(y, vf);
128 y += 16;
129 }
130 if XNN_UNLIKELY(elements != 0) {
131 // Prepare mask for valid 32-bit elements (depends on elements).
132 elements >>= 2 /* log2(sizeof(float)) */;
133 const __mmask16 vmask = _cvtu32_mask16((uint16_t) ((uint32_t) (UINT32_C(1) << elements) - UINT32_C(1)));
134
135 // Load up to 15 inputs at a time.
136 const __m512 vx = _mm512_maskz_loadu_ps(vmask, x);
137
138 // Compute reduced argument elements := round(x / log(2)).
139 const __m512 vn = _mm512_roundscale_ps(_mm512_mul_ps(vx, vlog2e), 0);
140
141 // Compute reduced argument t := x - elements * log(2).
142 // Use Cody-Waite range reduction method (note two constants to represent log(2)) to improve accuracy.
143 __m512 vt = _mm512_fmadd_ps(vn, vminus_ln2_hi, vx);
144 vt = _mm512_fmadd_ps(vn, vminus_ln2_lo, vt);
145
146 // Compute degree-5 polynomial approximation for exp(t) on [-log(2)/2, log(2)/2].
147 __m512 vp = _mm512_fmadd_ps(vc5, vt, vc4);
148 vp = _mm512_fmadd_ps(vp, vt, vc3);
149 vp = _mm512_fmadd_ps(vp, vt, vc2);
150 vp = _mm512_fmadd_ps(vp, vt, vc1);
151 vp = _mm512_fmadd_ps(vp, vt, vc0);
152
153 // Multiply "extended" floating-point numbers in ("mantissa", "exponent") representation.
154 __m512 vf = _mm512_mul_ps(vp, vscalev);
155 const __m512 ve = _mm512_add_ps(vn, vscalee);
156
157 // Multiply "mantissa" by the exp2("exponent").
158 vf = _mm512_scalef_ps(vf, ve);
159
160 // Store up to 15 results at a time.
161 _mm512_mask_storeu_ps(y, vmask, vf);
162 }
163 _mm256_zeroupper();
164 }
165