[petsc-dev] [Ideas-team] Seeking OLCF users complaining about poor build times

[Satish Balay]

On Fri, 27 Feb 2015, Victor Eijkhout wrote:

> 
> > On Feb 27, 2015, at 10:58 AM, Satish Balay <balay at mcs.anl.gov> wrote:
> > 
> >> Why would it be prohibited?  Especially if you remove the build after it's
> >> done?  What's a "simple run" and why do the security settings prohibit it?
> > 
> > I meant - "when configure creates and runs binaries - as part of configure step in /tmp".
> 
> That's prohibited at TACC too; we enforce that by removing some library on the login nodes. 
> The petsc build thankfully seems not to run into this.

There are 2 cases here.

1. can't run [mpi binaries] from front-end nodes. Here you have to use
--with-batch=1 anyway [and run the binary created by configure on
compute node] - so the rest of configure process happily uses /tmp for
intermediate files.

2. binaries would run from /home - but not /tmp [i.e /tmp is mounted
with noexec atribute]. This is the more restricted mode - where using
/tmp by configure is causing extra grief [then it would be when using
$PWD] - and this is the rare case I was refering to.

Satish

> The users can get around this restriction by downloading their own mpi, but mostly I'm guessing
> they use the mpi compilers from the environment.
> 
> Victor.
> 
> 
>