Re: [Starpu-devel] StarPU will cope by trying to purge

[Samuel Thibault <samuel.thibault@inria.fr>]

Hello,

Kadir Akbudak, on jeu. 22 juin 2017 10:12:36 +0300, wrote:
> I get lots of the following error and my program stalls forever:
> [starpu][_starpu_mktemp] Could not create temporary file in directory
> '/tmp/448/455', mskostemp failed with error 'Invalid argument'
> 
> The reason may be that the compute nodes of the Cray XC40 system that
> I use may not have /tmp. In fact compute nodes do not have disks at
> all.

Ok, so OOC is really not an option.

> I will investigate throttling task submission with the
> STARPU_LIMIT_MAX_SUBMITTED_TASKS environment variable.

That's currently the only way to limit the amount of data allocation for
MPI receives, yes.

> Regarding the shared memory system with 256GiB, I only run unix top
> program so I might have missed the spikes in memory consumption. What
> is the best way to show real memory consumption of StarPU?

The paje.trace file as opened in vite shows memory use in the MEMMANAGER
line.

> I tried to use starpu_data_display_memory_stats(), but it prints only
> "Memory stats :" and nothing else.

That indeed only displays some details on the registered data handles,
not global memory consumption.

Samuel