Is your application CPU bound or does it use other resources (disk, memory, network)? Sounds like it's not CPU bound. If your application is not CPU bound you're going to see poorer performance on a box with higher density cores running a job per core because there's more contention for disk and memory and network resources. On your dual core box you have at most two threads active vyying for these resources. On your 8 core box you have 8 threads, same number of disks and memory interfaces.- Ian
From: condor-users-bounces@xxxxxxxxxxx [mailto:condor-users-bounces@xxxxxxxxxxx] On Behalf Of Maryam Khademi
Sent: Thursday, December 04, 2008 10:32 PM
To: Condor-Users Mail List
Subject: [Condor-users] condor performance on quad core vs dual core machines
I have a performance question. Currently, I am using two clusters to execute my DAGs on.
On the first cluster, each machine has a two dual core CPU, whilst a machine on the second cluster has a two quad core CPU.
I run the same DAG on each cluster while there are equal number of cores available on both. For example, 8 unclaimed cores (2 machines) on cluster #1 and 8 unclaimed cores (1 machine) on cluster #2.
Interestingly, almost always I got better result on the dual core machines cluster!
These are the rest of particulars:
2 Dual Core
CPU: Xeon 3GHz
Condor Version: 6.8.0
Globus version: 4.0.4
2 Quad Core
CPU: Xeon 3GHz
OS: Red Hat Enterprise Linux Server release 5.1
OCS Version: Kusu 5 (build 1619)
Globus version: 4.2.1
P.S. All the above said are with the assumption that condor uses multi cores not single core in case of their availability.
This message may contain information that is confidential or otherwise protected from disclosure. If you are not the intended recipient, you are hereby notified that any use, disclosure, dissemination, distribution, or copying of this message, or any attachments, is strictly prohibited. If you have received this message in error, please advise the sender by reply e-mail, and delete the message and any attachments. Thank you.
Condor-users mailing list
To unsubscribe, send a message to condor-users-request@xxxxxxxxxxx with a
You can also unsubscribe by visiting
The archives can be found at: