0
I have a situtation where a Eclipse Workspace (identical) performs very differently on two very similar PCs. Both are approx 3-year old PCS, one i7-3520M, the other i5 (I do not have the exact model, but they are from 2013). Both have ssd drives. The i7 is a HP EliteBook, the i5 is a Lenovo Thinkpad... Both with 8G RAM... Both run Win7...
I can observe significant performance differences when checking out large SVN Repositories (lots of small file writes), and when starting a Jetty Server with a big Spring Context (lots of small file reads).
A process that takes 70 seconds on the i5, can take up to 160 seconds on the i7. CPU utilization is never at 100%, and swapping should not be the issue either.
I have managed to improve speed on the i7 by killing some processes (tivoli storage manager + cscript.exe) and by disabling antivirus, but it is still well over 100 seconds (so, still a remarkable difference). Besides, the i5 makes the 70 seconds with antivirus active...
So, knowing that the two machines have quite different configurations (different antivirus, different corporate applications, etc.), I want to find out what is slowing down the theoretically faster i7.
What can I do to find bottlenecks? Is there any tooling to diagnose this kind of issues? I can hardly imagine that it is a hardware related issue...
what is your hardware resource allocation on both systems when running the task? how much of that allocation is directly related to the operation you are profiling? In general IBM tools are often slow. its just the nature of java. – Frank Thomas – 2016-04-15T12:10:07.233
I don't see any bottlenecks, even the Harddrive seems to perform properly. I now found that the Antivirus is causing part of the slow performance, but there is still a difference. What I would expect as an answer is how to find this kind of bottlenecks in a methodic way (the Antivirus was a (logical) guess)... – Martin – 2016-04-21T13:13:02.403