Thank you Szillárd, that will indeed affect me and other users at the
cluster.
Best regards,
João
On Thu, Dec 5, 2013 at 8:08 AM, Szilárd Páll wrote:
> The first update on the redmine page does describe the source of the
> issue - although perhaps not in the most user-friendly manner.
>
> Tran
The first update on the redmine page does describe the source of the
issue - although perhaps not in the most user-friendly manner.
Translating it to a simple example, if you have dual-socket 8c nodes
and 2 GPUs per node:
mpirun -np 4 mdrun -ntomp 4 -gpu_id 0011 #single node 4 ranks per
node, 2-w
I did read it and wasn't very enlightened by it, that's why I kindly
requested for further information on the subject. "under certain
circumstances (but quite often), a segmentation fault occurs" definitely
sounds serious, but it's not very explicit.
In any case, I'll request a new install and tel
On Wed, Dec 4, 2013 at 10:15 AM, João Henriques
wrote:
> Soon enough we will have daily releases :P
I hope you're not suggesting that we should release less frequently! :)
> Mark, can you please elaborate just a tiny bit longer on how relevant were
> the "GPU-load balancing and GPU-sharing bugs
Soon enough we will have daily releases :P
Mark, can you please elaborate just a tiny bit longer on how relevant were
the "GPU-load balancing and GPU-sharing bugs introduced in a bug fix in
4.6.4"? I'm asking this because I just requested the support team at my
local cluster to install 4.6.4 and I
Hi GROMACS users,
GROMACS 4.6.5 is officially released. It contains a significant bug fix
when using GPUs in a multi-node simulation, introduced in another recent
bug fix. We encourage all users to upgrade their installations from earlier
4.6.x releases.
You can find the code, manual, release not