eden-emu/opus
15
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Support for plain AVX with no FMA

Browse source
This commit is contained in:
Jean-Marc Valin 2018-12-02 02:10:45 -05:00
parent 91d90676e1
commit 771cc7868a
1 changed files with 42 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

44
dnn/nnet.c
View file

@ -41,8 +41,11 @@
#define SOFTMAX_HACK #define SOFTMAX_HACK
#ifdef __AVX2__ #ifdef __AVX__
#include <immintrin.h> #include <immintrin.h>
#ifdef __AVX2__
static __m256 exp8_approx(__m256 X) static __m256 exp8_approx(__m256 X)
{ {
const __m256 K0 = _mm256_set1_ps(0.99992522f); const __m256 K0 = _mm256_set1_ps(0.99992522f);
@ -65,7 +68,44 @@ static __m256 exp8_approx(__m256 X)
Y = _mm256_castsi256_ps(_mm256_and_si256(mask, _mm256_add_epi32(I, _mm256_castps_si256(Y)))); Y = _mm256_castsi256_ps(_mm256_and_si256(mask, _mm256_add_epi32(I, _mm256_castps_si256(Y))));
return Y; return Y;
} }
#else
#define _mm256_fmadd_ps(a,b,c) _mm256_add_ps(_mm256_mul_ps(a, b), c)
#define _mm_fmadd_ps(a,b,c) _mm_add_ps(_mm_mul_ps(a, b), c)
static __m128 exp4_approx(__m128 X)
{
const __m128 K0 = _mm_set1_ps(0.99992522f);
const __m128 K1 = _mm_set1_ps(0.69583354f);
const __m128 K2 = _mm_set1_ps(0.22606716f);
const __m128 K3 = _mm_set1_ps(0.078024523f);
const __m128 log2_E = _mm_set1_ps(1.44269504);
const __m128 max_in = _mm_set1_ps(50.f);
const __m128 min_in = _mm_set1_ps(-50.f);
const __m128i mask = _mm_set1_epi32(0x7fffffff);
__m128 XF, Y;
__m128i I;
X = _mm_mul_ps(X, log2_E);
X = _mm_max_ps(min_in, _mm_min_ps(max_in, X));
XF = _mm_floor_ps(X);
I = _mm_cvtps_epi32(XF);
X = _mm_sub_ps(X, XF);
Y = _mm_fmadd_ps(_mm_fmadd_ps(_mm_fmadd_ps(K3, X, K2), X, K1), X, K0);
I = _mm_slli_epi32(I, 23);
Y = _mm_castsi128_ps(_mm_and_si128(mask, _mm_add_epi32(I, _mm_castps_si128(Y))));
return Y;
}
static __m256 exp8_approx(__m256 X)
{
__m256 Y;
__m128 Xhi, Xlo, Yhi, Ylo;
Xhi = _mm256_extractf128_ps(X, 1);
Xlo = _mm256_extractf128_ps(X, 0);
Yhi = exp4_approx(Xhi);
Ylo = exp4_approx(Xlo);
Y = _mm256_insertf128_ps(_mm256_setzero_ps(), Yhi, 1);
Y = _mm256_insertf128_ps(Y, Ylo, 0);
return Y;
}
#endif
static float celt_exp(float x) static float celt_exp(float x)
{ {