99c7bba8e4
This patch adds the basic infrastructure for adding the SkylakeX (Intel Skylake server) target. The SkylakeX target will use the AVX512 (AVX512VL level) instruction set, which brings 2 basic things: 1) 512 bit wide SIMD (2x width of AVX2) 2) 32 SIMD registers (2x the number on AVX2) This initial patch only contains a trivial transofrmation of the Haswell SGEMM kernel to AVX512VL; more will follow later but this patch aims to get the infrastructure in place for this "later". Full performance tuning has not been done yet; with more registers and wider SIMD it's in theory possible to retune the kernels but even without that there's an interesting enough performance increase (30-40% range) with just this change.
139 lines
3.9 KiB
C
139 lines
3.9 KiB
C
/***************************************************************************
|
|
Copyright (c) 2014, The OpenBLAS Project
|
|
All rights reserved.
|
|
Redistribution and use in source and binary forms, with or without
|
|
modification, are permitted provided that the following conditions are
|
|
met:
|
|
1. Redistributions of source code must retain the above copyright
|
|
notice, this list of conditions and the following disclaimer.
|
|
2. Redistributions in binary form must reproduce the above copyright
|
|
notice, this list of conditions and the following disclaimer in
|
|
the documentation and/or other materials provided with the
|
|
distribution.
|
|
3. Neither the name of the OpenBLAS project nor the names of
|
|
its contributors may be used to endorse or promote products
|
|
derived from this software without specific prior written permission.
|
|
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
|
|
AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
|
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
|
ARE DISCLAIMED. IN NO EVENT SHALL THE OPENBLAS PROJECT OR CONTRIBUTORS BE
|
|
LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
|
|
DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
|
|
SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
|
|
CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
|
|
OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE
|
|
USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
|
*****************************************************************************/
|
|
|
|
|
|
#include "common.h"
|
|
|
|
|
|
#if defined(PILEDRIVER) || defined(STEAMROLLER) || defined(EXCAVATOR)
|
|
#include "caxpy_microk_steamroller-2.c"
|
|
#elif defined(BULLDOZER)
|
|
#include "caxpy_microk_bulldozer-2.c"
|
|
#elif defined(HASWELL) || defined(ZEN) || defined(SKYLAKEX)
|
|
#include "caxpy_microk_haswell-2.c"
|
|
#elif defined(SANDYBRIDGE)
|
|
#include "caxpy_microk_sandy-2.c"
|
|
#endif
|
|
|
|
|
|
#ifndef HAVE_KERNEL_8
|
|
|
|
static void caxpy_kernel_8(BLASLONG n, FLOAT *x, FLOAT *y, FLOAT *alpha)
|
|
{
|
|
BLASLONG register i = 0;
|
|
BLASLONG register ix = 0;
|
|
FLOAT da_r = alpha[0];
|
|
FLOAT da_i = alpha[1];
|
|
|
|
|
|
while(i < n)
|
|
{
|
|
#if !defined(CONJ)
|
|
y[ix] += ( da_r * x[ix] - da_i * x[ix+1] ) ;
|
|
y[ix+1] += ( da_r * x[ix+1] + da_i * x[ix] ) ;
|
|
y[ix+2] += ( da_r * x[ix+2] - da_i * x[ix+3] ) ;
|
|
y[ix+3] += ( da_r * x[ix+3] + da_i * x[ix+2] ) ;
|
|
#else
|
|
y[ix] += ( da_r * x[ix] + da_i * x[ix+1] ) ;
|
|
y[ix+1] -= ( da_r * x[ix+1] - da_i * x[ix] ) ;
|
|
y[ix+2] += ( da_r * x[ix+2] + da_i * x[ix+3] ) ;
|
|
y[ix+3] -= ( da_r * x[ix+3] - da_i * x[ix+2] ) ;
|
|
#endif
|
|
|
|
ix+=4 ;
|
|
i+=2 ;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
int CNAME(BLASLONG n, BLASLONG dummy0, BLASLONG dummy1, FLOAT da_r, FLOAT da_i, FLOAT *x, BLASLONG inc_x, FLOAT *y, BLASLONG inc_y, FLOAT *dummy, BLASLONG dummy2)
|
|
{
|
|
BLASLONG i=0;
|
|
BLASLONG ix=0,iy=0;
|
|
FLOAT da[2];
|
|
|
|
if ( n <= 0 ) return(0);
|
|
|
|
if ( (inc_x == 1) && (inc_y == 1) )
|
|
{
|
|
|
|
BLASLONG n1 = n & -32;
|
|
|
|
if ( n1 )
|
|
{
|
|
da[0] = da_r;
|
|
da[1] = da_i;
|
|
caxpy_kernel_8(n1, x, y , da );
|
|
ix = 2 * n1;
|
|
}
|
|
|
|
i = n1;
|
|
while(i < n)
|
|
{
|
|
#if !defined(CONJ)
|
|
y[ix] += ( da_r * x[ix] - da_i * x[ix+1] ) ;
|
|
y[ix+1] += ( da_r * x[ix+1] + da_i * x[ix] ) ;
|
|
#else
|
|
y[ix] += ( da_r * x[ix] + da_i * x[ix+1] ) ;
|
|
y[ix+1] -= ( da_r * x[ix+1] - da_i * x[ix] ) ;
|
|
#endif
|
|
i++ ;
|
|
ix += 2;
|
|
|
|
}
|
|
return(0);
|
|
|
|
|
|
}
|
|
|
|
inc_x *=2;
|
|
inc_y *=2;
|
|
|
|
while(i < n)
|
|
{
|
|
|
|
#if !defined(CONJ)
|
|
y[iy] += ( da_r * x[ix] - da_i * x[ix+1] ) ;
|
|
y[iy+1] += ( da_r * x[ix+1] + da_i * x[ix] ) ;
|
|
#else
|
|
y[iy] += ( da_r * x[ix] + da_i * x[ix+1] ) ;
|
|
y[iy+1] -= ( da_r * x[ix+1] - da_i * x[ix] ) ;
|
|
#endif
|
|
ix += inc_x ;
|
|
iy += inc_y ;
|
|
i++ ;
|
|
|
|
}
|
|
return(0);
|
|
|
|
}
|
|
|
|
|