[Cado-nfs-discuss] [error] on linear algebra process, cannot allocate memory but there's more info to that
Emmanuel.Thome at inria.fr
Wed Aug 29 11:51:40 CEST 2018
On Wed, Aug 22, 2018 at 11:00:30AM +0200, Paul Zimmermann wrote:
> > Hi, i have encountered another problem after the sieving and filtering step.
> > Have done some parameter tuning at first for lpb0 and lpb1 increased by 1
> > from default parameter in c120 but encounter this error:
> > [...]
> > /home/chunnie/Desktop/Math3/cado-nfs/build/ubuntu/filter/replay -purged
> > /tmp/cado.uf94w5gq/c110.purged.gz -his /tmp/cado.uf94w5gq/c110.history.gz
> > -index /tmp/cado.uf94w5gq/c110.index.gz -out
> > /tmp/cado.uf94w5gq/c110.sparse.bin > /tmp/cado.uf94w5gq/c110.replay.stdout.2
> > 2> /tmp/cado.uf94w5gq/c110.replay.stderr.2
> > Error:Filtering - Merging: Stderr output follows (stored in file
> > /tmp/cado.uf94w5gq/c110.replay.stderr.2):
> > b'antebuffer set to
> > /home/chunnie/Desktop/Math3/cado-nfs/build/ubuntu/utils/antebuffer\nfork:
> > Cannot allocate memory\ncode BUG() : condition indexfile != NULL failed in
> > writeIndex at /home/chunnie/Desktop/Math3/cado-nfs/filter/replay.c:560 --
> > Abort\n'
> > [...]
> > is the only solution for this is to enlarge memory? or is there any parameter
> > that i need to tune again for this to work?
> it seems you cannot even allocate the few bytes necessary to open a file
> descriptor to write into the index file, thus indeed you should enlarge memory.
I'd rather bet on the filesystem in /tmp being full, in that case.
We could/should report the errors more nicely.
> (You can also try to reproduce the error by running by hand the full replay
> command above.)
> Paul Zimmermann
> Cado-nfs-discuss mailing list
> Cado-nfs-discuss at lists.gforge.inria.fr
More information about the Cado-nfs-discuss