Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

We have two servers that are very similar configurations, but when running the same custom in-house application, the performance numbers are substantially different.

Here are the server specs Server1 Dell R815, opteron 6168, 1.9 GHz, 64GB DDR3 1333 ram, 48 cores, Ubuntu 10.10

Server2 Dell R815, opteron 6238, 2.6 GHz, 128GB DDR3 1600 ram, 48 cores Ubuntu 10.4.4 LTS

Both servers have identical bios settings with node interleaving off, and all power related settings on max performance. NUMA is turned off on both machines, and both have base installations with only the packages needed for NFS, SSH, monitoring utilities, and our application. Both have the same version of our in-house application, and it is configured the same on both machines. Both machines have all cores at or near 100% all of the time.

The strange thing is server2 is only performing about 20% of the work that server1 is! I have read about a large increase in context switching latency between the 2.6 and 3.2 kernels, but that doesn't apply in this case.

Question: How can I troubleshoot this performance discrepancy between the two systems?


share|improve this question
Since Server 2 is much beefier than Server 1, it logically follows that it can complete a given amount of work in less time. – Michael Hampton Mar 15 '13 at 15:42
Please note that server2 is the slow performer. – user2174608 Mar 15 '13 at 16:43

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.