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

Satish Balay balay at mcs.anl.gov
Thu Feb 26 21:26:11 CST 2015


/tmp is usually a local [to the machine filesyste] - or tmpfs [i.e
using RAM]. The following command usually shows the filesystem type.

[tmpfs over RAM]
$ df -T /tmp
Filesystem     Type  1K-blocks  Used Available Use% Mounted on
tmpfs          tmpfs   3935432  8784   3926648   1% /tmp

[local disk]
$ df -T /tmp
Filesystem    Type   1K-blocks      Used Available Use% Mounted on
/dev/sda1     ext3     9614116   6642292   2483452  73% /

But I suspect Barry is refering to TMPDIR [when he said /tmp]

On tital TMPDIR is /tmp. [per configure.log]

On hopper - the default is not /tmp

balay at hopper01:~/petsc.clone> echo $TMPDIR
/scratch/scratchdirs/balay
balay at hopper01:~/petsc.clone> df -T /scratch/scratchdirs/balay
Filesystem                               Type       1K-blocks        Used     Available Use% Mounted on
10.10.10.1 at o2ib:10.10.10.2 at o2ib:/scratch lustre 1208581506288 77197515368 1070259709628   7% /scratch
balay at hopper01:~/petsc.clone> 

Satish

On Thu, 26 Feb 2015, Nathan Collier wrote:

> Barry,
> 
> > Do you know where /tmp is on the system? Presumably it is fast?
> 
> I am not sure about that. There is a /tmp but I am not sure what it is or
> if/how I can use it.
> 
> Also, I was the only one running anything on the login node when I did the
> configure/make. My configure does include the sowing download.
> 
> Nate
> 
> 
> On Thu, Feb 26, 2015 at 9:53 PM, Satish Balay <balay at mcs.anl.gov> wrote:
> 
> > On Thu, 26 Feb 2015, Nathan Collier wrote:
> >
> > > Ok, so I am not sure I got all that, but here goes some more information
> > > (attached log).
> > >
> > > No packages / master
> > >
> > > [nate at titan] time python no_packages.py
> > > ...
> > > real 14m32.169s
> > > user 1m51.527s
> > > sys 3m40.734s
> > >
> > > [nate at titan] time make all
> > > ...
> > > real 15m56.004s
> > > user 8m8.971s
> > > sys 52m42.734s
> >
> > BTW: This number for 'make all' looks bizzare [or I don't know how to
> > interpret it].
> >
> > - 'sys' time is 52m?
> > - 'user' time is 8m - but 'real' is 15m - so its not a parallel build?
> >    But configure sets 'MAKE_NP = 24' (with gnumake build).
> >
> > Perhaps its all intel compiler license server overhead? :)
> >
> > Satish
> >
> 




More information about the petsc-dev mailing list