Charm++ parallel programming system

[Evan Ramos <evan AT hpccharm.com>]

Thanks for the quick reply! This output suggests that configure.ac's code to generate conv-mach-opt.{sh,mak} is not functioning properly. This may come down to differences between sed variants. I am spinning up an Alpine Linux VM to debug this.

--

Evan A. Ramos

Software Engineer

Charmworks, Inc.

On Mon, Oct 29, 2018 at 5:29 PM Jozsef Bakosi <jbakosi AT lanl.gov> wrote:

On 10.29.2018 17:26, Evan Ramos wrote:
> If you run `grep -i charmdebug config.log charmconfig.out conv-mach-opt.*`,
> what is printed?

/home/quinoa/tpl/gnu/charm/mpi-linux-x86_64/tmp # grep -i charmdebug config.log charmconfig.out conv-mach-opt.*
config.log:  $ ./configure --disable-controlpoint --disable-tracing --disable-tracing-commthread --disable-charmdebug --disable-replay --disable-error-checking --disable-stats --enable-error-checking
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:| #define CMK_CHARMDEBUG 0
config.log:#define CMK_CHARMDEBUG 0
charmconfig.out:CharmDebug is disabled


>
> --
> Evan A. Ramos
> Software Engineer
> Charmworks, Inc.
>
>
> On Mon, Oct 29, 2018 at 4:45 PM Jozsef Bakosi <jbakosi AT lanl.gov> wrote:
>
> > On 10.29.2018 12:38, Evan Ramos wrote:
> > > Hi Jozsef,
> > >
> > > I believe this patch should fix the __malloc_hook problem:
> > > https://charm.cs.illinois.edu/gerrit/c/charm/+/4744
> >
> > Applying the patch and re-building with
> >
> > $ ./build charm++ mpi-linux-x86_64 --enable-error-checking --build-shared
> > --with-production -j36 -w -fpermissive -O3 -DNDEBUG
> >
> > I get, in tmp/config.log:
> >
> >   $ ./configure --disable-controlpoint --disable-tracing
> > --disable-tracing-commthread --disable-charmdebug --disable-replay
> > --disable-error-checking --disable-stats --enable-error-checking
> >
> > and the error:
> >
> > In file included from memory.C:677:
> > memory-charmdebug.c:30:2: error: #error "charmdebug is not enabled (e.g.
> > when building with-production)"
> >  #error "charmdebug is not enabled (e.g. when building with-production)"
> >   ^~~~~
> >
> > I guess, charmdebug is disabled even without me passing
> > --disable-charmdebug to it, so that turns out to be a problem.
> >
> > > I am confused by the charmdebug error. I looked at the build system and
> > > don't see how memory-charmdebug would be included from any file if
> > > --disable-charmdebug was passed. Does the configure script say
> > "CharmDebug
> > > is disabled"?
> >
> > Here are the first few lines of tmp/config.log for when I explicitly
> > disable charmdebug:
> >
> > ==========
> > This file contains any messages produced by compilers while
> > running configure, to aid debugging if configure makes a mistake.
> >
> > It was created by Charm++ configure 60900, which was
> > generated by GNU Autoconf 2.69.  Invocation command line was
> >
> >   $ ./configure --disable-controlpoint --disable-tracing
> > --disable-tracing-commthread --disable-charmdebug --disable-replay
> > --disable-error-checking --disable-stats --enable-error-checking
> > --disable-charmdebug
> > ==========
> >
> > I guess --disable-charmdebug is triggered by --with-production, so both
> > will do the same and for some reason this breaks on alpine.
> >
> > Thanks,
> > Jozsef
> >