cblas.h defines
double cblas_dcabs1(const void *z);
float cblas_scabs1(const void *c);
but does not provide an implementation. This commit adds
the missing implementation in the common CBLAS style and
eliminates the linker error when calling the above function.
long's in windows are 4 bytes (MSVS, intel compilers). Use int64_t and int32_t
to ensure 8 byte integers for ILP interface.
support 8 byte integer flag for intel ifort compiler
Commits 2d9fbdeecc and
628a2095c2 changed the function prototypes in the
cblas_f77.h.
However, the F77_xerbla case was not treated in the same way as the others. In
particular, the consequence is now that the xerbla function provided by the
CBLAS/src/xerbla.c file is never called at runtime (because it is identified as
“F77_xerbla” in the symbol table, instead of simply “xerbla”, since F77_xerbla
is *not* a macro).
This patch restores the symmetry between xerbla and the other functions.
Ancient CMake versions required upper-case commands. Later command
names became case-insensitive. Now the preferred style is lower-case.
Run the following shell code:
NOTE: MUST USE GNU compliant version of sed
cmake --help-command-list \
| grep -v "cmake version" \
| while read c; do
echo 's/\b'"$(echo $c | tr '[:lower:]' '[:upper:]')"'\(\s*\)(/'"$c"'\1(/g'
done >convert.sed \
&& git ls-files -z -- bootstrap '*.cmake' '*.cmake.in' '*CMakeLists.txt' \
| xargs -0 gsed -i -f convert.sed \
&& rm convert.sed
This is mostly a long term maintenance improvement.
Many coding styles require elimination of trailing whitespace, and
many editors and source code management configurations automatically
gobble up whitespace. When these tools gobble up whitespace, it
complicates reviewing the meaningful code changes.
By removing whitespace on one patch, it makes future
code reviews much easier.
=SCRIPT====================================================================
if which tempfile &>/dev/null; then
TEMPMAKER=tempfile
elif which mktemp &>/dev/null; then
TEMPMAKER=mktemp
else
echo "Cannot find tempfile program." 2>&1
exit 1
fi
MYTEMP=$($TEMPMAKER)
trap 'rm -f $MYTEMP' SIGINT SIGTERM
stripit() {
echo "stripping $1"
sed 's/[ \t]*$//' "$1" > $MYTEMP
cp $MYTEMP "$1"
}
if [ $# -gt 0 ]; then
while [ "$1" != "" ]; do
stripit $1
shift
done
else
while read -t 2; do
stripit $REPLY
done
fi
rm $MYTEMP
=================================================
Use CMake capabilities to test if the specified fortran
and C compilers are compatible. Fail early if they are
determined to be incompatible.
Use CMake capabilities to generate platform specific
header files, If CMake FortranCInterface_HEADER function
fails to identify FortranCInterface_GLOBAL_FOUND or
FortranCInterface_MODULE_FOUND, then fall back to the
statically generated file that is used by Makefiles.
The cblas_mangling_with_flags.h and lapacke_mangling_with_flags.h
are template files that are not used directly. They need to be
configured (i.e. copied) to lapacke_mangling.h and cblas_mangling.h
header files that are used. These renamings make the intent of
these files more clearly reported.
The file LAPACKE/include/lapacke_mangling.h should not be stored in the
repository because it should be generated (copied from above) at build
time.
Netlib CBLAS can be put on top of BLAS with 64-bit integers.
It can be accomplished using the flag "WeirdNEC" :)
It basically redefines int to long, but at the same time, it had multiple bugs.