Hello.
We have a situation where it looks like SEP is ignoring our exclusion filters.
After migrating to a new environment we found that a stress test program we use for testing our software would run slowly. In the old environment the test would finish after several hours. In the new environment, we let the test run for a couple days and it was not able to finish.
This is odd because the new environment has much better hardware. We expected the test to finish more quickly. Using the free profiler from AMD we found that a bunch of time was being spent in these modules from SEP:
BHDDrvx64.sys
UMEngx86.dll
This was also odd because we have an exclusion for the directory where the tester executable, 'testee', and the logs that they generate reside.
We found that uninstalling SEP gave us the performance we expected ( i.e. better than in the old environment ). We also found that if we turned off logging in our executables the performance was more like we expected.
So it seems that our performance problem comes down to SEP not using our exclusion filter. D:\builds\*
Below our descriptions of our old and new environments
- Old environment:
OS: Windows 2003
SEP: 12.1.4013.4013 (32-bit)
- New environment
OS: Windows 2012 R2
SEP: 12.1.4013.4013 (64 bit)
Looking for any ideas or suggestions.
Thanks.
Jason