Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Thu, 8 Aug 2013 01:14:15 +0200
From: Lukas Odzioba <>
Subject: Re: --fork using different OpenCL devices

2013/8/8 magnum <>:
> The faster device would run at full speed but that process will complete earlier because the job would be split in half. This is the same problem as when running MPI with heterogenous nodes. It is possible to mitigate using some option that internally converts to -node functionality, ie. the slower device is node 1 of 3 and the faster one is nodes 2 and 3 of 3. The only hard part in implementing that would be what option syntax to use...

Can we measure when each "node" finished their work? With this we
could divide workloads based on time needed to compute - self adapting
method that should work well even if you change frequency of gpus
after session starts. It would be cool if we could avoid hardcoding
those values or having user to choose them.


Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.