Skip to content
GitLab
Commit 276c4948 authored by Emmanuel Bertin's avatar Emmanuel Bertin
Browse files

Added support for CORE-AVX2 and CORE-AVC-I optimized code when using INTEL compiler. 

Added better control of the number of threads used in MKL library calls (when used).
Pushed version number to 2.18.3.
parent 79250284
Hide whitespace changes
Inline Side-by-side
acx_prog_cc_optim.m4
View file @ 276c4948
......@@ -25,7 +25,7 @@ dnl You should have received a copy of the GNU General Public License
dnl along with AstrOmatic software.
dnl If not, see <http://www.gnu.org/licenses/>.
dnl
dnl Last modified: 05/07/2012
dnl Last modified: 19/07/2012
dnl
dnl %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
dnl
......@@ -60,15 +60,15 @@ EOF
if $CC -V 2>&1 | grep -i "Intel(R) 64" > /dev/null 2>&1 &&
$CC -c -O conftest.c > /dev/null 2>&1 &&
test -f conftest.o; then
prog_cc_optim_cv_flags="-O3 -axSSE3,SSE4.1,SSE4.2,AVX -no-prec-div -unroll"
prog_ld_optim_cv_flags=""
prog_cc_optim_cv_flags="-O3 -axSSE3,SSE4.1,SSE4.2,AVX,CORE-AVX2,CORE-AVX-I -no-prec-div -unroll"
prog_ld_optim_cv_flags="-static-intel"
dnl INTEL C 32bits compiler
elif $CC -V 2>&1 | grep -i "Intel(R)" > /dev/null 2>&1 &&
$CC -c -O conftest.c > /dev/null 2>&1 &&
test -f conftest.o; then
prog_cc_optim_cv_flags="-O3 -axSSE2,SSE3,SSE4.1,SSE4.2,AVX -no-prec-div -unroll"
prog_ld_optim_cv_flags=""
prog_cc_optim_cv_flags="-O3 -axSSE2,SSE3,SSE4.1,SSE4.2,AVX,CORE-AVX2,CORE-AVX-I -no-prec-div -unroll"
prog_ld_optim_cv_flags="-static-intel"
dnl GCC
elif test "$GCC" = "yes"; then
......@@ -129,13 +129,10 @@ EOF
rm -f conftest.*
fi
if test -n "$prog_cc_optim_cv_flags"; then
AM_CFLAGS="$AM_CFLAGS $prog_cc_optim_cv_flags"
AM_CFLAGS="$CFLAGS $prog_cc_optim_cv_flags"
AM_LDFLAGS="$LDFLAGS $prog_ld_optim_cv_flags"
else
prog_cc_optim_cv_flags=""
fi
if test -n "$prog_ld_optim_cv_flags"; then
AM_LDFLAGS="$AM_LDFLAGS $prog_ld_optim_cv_flags"
else
prog_ld_optim_cv_flags=""
fi
])
......
configure
View file @ 276c4948
#! /bin/sh
# Guess values for system-dependent variables and create Makefiles.
# Generated by GNU Autoconf 2.68 for sextractor 2.18.2.
# Generated by GNU Autoconf 2.68 for sextractor 2.18.3.
#
# Report bugs to <bertin@iap.fr>.
#
......@@ -570,8 +570,8 @@ MAKEFLAGS=
# Identity of this package.
PACKAGE_NAME='sextractor'
PACKAGE_TARNAME='sextractor'
PACKAGE_VERSION='2.18.2'
PACKAGE_STRING='sextractor 2.18.2'
PACKAGE_VERSION='2.18.3'
PACKAGE_STRING='sextractor 2.18.3'
PACKAGE_BUGREPORT='bertin@iap.fr'
PACKAGE_URL=''
......@@ -1324,7 +1324,7 @@ if test "$ac_init_help" = "long"; then
# Omit some internal or obsolete options to make the list less imposing.
# This message is too long to be a string in the A/UX 3.1 sh.
cat <<_ACEOF
\`configure' configures sextractor 2.18.2 to adapt to many kinds of systems.
\`configure' configures sextractor 2.18.3 to adapt to many kinds of systems.
Usage: $0 [OPTION]... [VAR=VALUE]...
......@@ -1394,7 +1394,7 @@ fi
if test -n "$ac_init_help"; then
case $ac_init_help in
short | recursive ) echo "Configuration of sextractor 2.18.2:";;
short | recursive ) echo "Configuration of sextractor 2.18.3:";;
esac
cat <<\_ACEOF
......@@ -1524,7 +1524,7 @@ fi
test -n "$ac_init_help" && exit $ac_status
if $ac_init_version; then
cat <<\_ACEOF
sextractor configure 2.18.2
sextractor configure 2.18.3
generated by GNU Autoconf 2.68
Copyright (C) 2010 Free Software Foundation, Inc.
......@@ -1947,7 +1947,7 @@ cat >config.log <<_ACEOF
This file contains any messages produced by compilers while
running configure, to aid debugging if configure makes a mistake.
It was created by sextractor $as_me 2.18.2, which was
It was created by sextractor $as_me 2.18.3, which was
generated by GNU Autoconf 2.68. Invocation command line was
$ $0 $@
......@@ -2752,7 +2752,7 @@ fi
# Define the identity of the package.
PACKAGE='sextractor'
VERSION='2.18.2'
VERSION='2.18.3'
cat >>confdefs.h <<_ACEOF
......@@ -3948,14 +3948,14 @@ EOF
if $CC -V 2>&1 | grep -i "Intel(R) 64" > /dev/null 2>&1 &&
$CC -c -O conftest.c > /dev/null 2>&1 &&
test -f conftest.o; then
prog_cc_optim_cv_flags="-O3 -axSSE3,SSE4.1,SSE4.2,AVX -no-prec-div -unroll"
prog_ld_optim_cv_flags=""
prog_cc_optim_cv_flags="-O3 -axSSE3,SSE4.1,SSE4.2,AVX,CORE-AVX2,CORE-AVX-I -no-prec-div -unroll"
prog_ld_optim_cv_flags="-static-intel"
elif $CC -V 2>&1 | grep -i "Intel(R)" > /dev/null 2>&1 &&
$CC -c -O conftest.c > /dev/null 2>&1 &&
test -f conftest.o; then
prog_cc_optim_cv_flags="-O3 -axSSE2,SSE3,SSE4.1,SSE4.2,AVX -no-prec-div -unroll"
prog_ld_optim_cv_flags=""
prog_cc_optim_cv_flags="-O3 -axSSE2,SSE3,SSE4.1,SSE4.2,AVX,CORE-AVX2,CORE-AVX-I -no-prec-div -unroll"
prog_ld_optim_cv_flags="-static-intel"
elif test "$GCC" = "yes"; then
prog_cc_optim_cv_flags="-O3 -g -funroll-loops -fomit-frame-pointer -Wall"
......@@ -4008,13 +4008,10 @@ EOF
rm -f conftest.*
fi
if test -n "$prog_cc_optim_cv_flags"; then
AM_CFLAGS="$AM_CFLAGS $prog_cc_optim_cv_flags"
AM_CFLAGS="$CFLAGS $prog_cc_optim_cv_flags"
AM_LDFLAGS="$LDFLAGS $prog_ld_optim_cv_flags"
else
prog_cc_optim_cv_flags=""
fi
if test -n "$prog_ld_optim_cv_flags"; then
AM_LDFLAGS="$AM_LDFLAGS $prog_ld_optim_cv_flags"
else
prog_ld_optim_cv_flags=""
fi
......@@ -15512,7 +15509,7 @@ cat >>$CONFIG_STATUS <<\_ACEOF || ac_write_fail=1
# report actual input values of CONFIG_FILES etc. instead of their
# values after options handling.
ac_log="
This file was extended by sextractor $as_me 2.18.2, which was
This file was extended by sextractor $as_me 2.18.3, which was
generated by GNU Autoconf 2.68. Invocation command line was
CONFIG_FILES = $CONFIG_FILES
......@@ -15578,7 +15575,7 @@ _ACEOF
cat >>$CONFIG_STATUS <<_ACEOF || ac_write_fail=1
ac_cs_config="`$as_echo "$ac_configure_args" | sed 's/^ //; s/[\\""\`\$]/\\\\&/g'`"
ac_cs_version="\\
sextractor config.status 2.18.2
sextractor config.status 2.18.3
configured by $0, generated by GNU Autoconf 2.68,
with options \\"\$ac_cs_config\\"
......
configure.ac
View file @ 276c4948
......@@ -22,7 +22,7 @@
# You should have received a copy of the GNU General Public License
# along with SExtractor. If not, see <http://www.gnu.org/licenses/>.
#
# Last modified: 17/07/2012
# Last modified: 19/07/2012
#
#%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
......@@ -31,7 +31,7 @@ define([AC_CACHE_LOAD],)
define([AC_CACHE_SAVE],)
# This is your standard Bertin source code...
AC_INIT(sextractor, 2.18.2, [bertin@iap.fr])
AC_INIT(sextractor, 2.18.3, [bertin@iap.fr])
AC_CONFIG_SRCDIR(src/makeit.c)
AC_CONFIG_AUX_DIR(autoconf)
AC_CONFIG_HEADERS(config.h)
......
man/sex.1
View file @ 276c4948
.TH SEXTRACTOR "1" "July 2012" "SExtractor 2.18.2" "User Commands"
.TH SEXTRACTOR "1" "July 2012" "SExtractor 2.18.3" "User Commands"
.SH NAME
sex \- extract a source catalogue from an astronomical FITS image
.SH SYNOPSIS
......
src/fits/fitshead.c
View file @ 276c4948
......@@ -23,7 +23,7 @@
* along with AstrOmatic software.
* If not, see <http://www.gnu.org/licenses/>.
*
* Last modified: 03/06/2012
* Last modified: 18/07/2012
*
*%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%*/
......@@ -705,16 +705,17 @@ PURPOSE Give the ``t_type'' of a binary-table field from its ``TFORM''.
INPUT TFORM string (see the FITS documentation).
OUTPUT size in bytes, or RETURN_ERROR if the TFORM is unknown.
NOTES -.
AUTHOR E. Bertin (IAP & Leiden observatory)
VERSION 28/10/2009
AUTHOR E. Bertin (IAP)
VERSION 18/07/2012
***/
t_type ttypeof(char *str)
{
char *str2;
long dummy;
str2 = str;
strtol(str, &str2, 10);
dummy = strtol(str, &str2, 10);
switch ((int)*str2)
{
case 'L': case 'B': case 'X': return T_BYTE;
......
src/prefs.c
View file @ 276c4948
......@@ -456,10 +456,6 @@ void preprefs()
"Performance will be degraded.");
#endif
#ifdef HAVE_MKL
mkl_set_num_threads(prefs.nthreads);
#endif
#else
if (prefs.nthreads != 1)
{
......@@ -469,6 +465,11 @@ void preprefs()
}
#endif
/* We temporarily allow threads for non-multithreaded code (defaulted to 1)*/
#ifdef HAVE_MKL
mkl_set_num_threads(prefs.nthreads);
#endif
/* Override INTEL CPU detection routine to help performance on 3rd-party CPUs */
#if defined(__INTEL_COMPILER) && defined (USE_CPUREDISPATCH)
__get_cpuid(1, &eax, &ebx, &ecx, &edx);
......
Supports Markdown
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment