Unfortunately, this hasn't resolved our deadlock, but thanks!
When checking the deadlock on 4 localities, I found out that on 3 localities the processes consume 100% CPU while on one the process just idles with CPU usage close to 0%.
Does this somehow help to narrow down the cause? Thanks Tim On 02/23/2017 04:53 PM, Thomas Heller wrote:
On 02/23/2017 02:49 PM, Tim Biedert wrote:Thanks for the hint. We're already getting deadlocks at 4 localities and we're not explicitly using any barriers. Are there any internal barriers in HPX which might be affected by this just due to a high task count?No. However, while attempting to fix #2516, I discovered something bad that happens during parcel decoding. The fix I came up with is in the following PR: https://github.com/STEllAR-GROUP/hpx/pull/2518 This might also be the reason for the deadlock you are seeing. Could you please try this patch and see if that fixes your deadlocks?Best, Tim On 02/22/2017 10:30 PM, Thomas Heller wrote:If you are running with at least 256 localities (or are using hpx::lcos::barrier with more than 256 participating threads), please be aware of: https://github.com/STEllAR-GROUP/hpx/issues/2516_______________________________________________ hpx-users mailing list [email protected] https://mail.cct.lsu.edu/mailman/listinfo/hpx-users
smime.p7s
Description: S/MIME Cryptographic Signature
_______________________________________________ hpx-users mailing list [email protected] https://mail.cct.lsu.edu/mailman/listinfo/hpx-users
