Hallvard B Furuseth wrote:

Looking at liblutil/utils.c:lutil_gettime() led me to

Beware of QueryPerformanceCounter()
http://www.virtualdub.org/blog/pivot/entry.php?id=106

Is warning relevant to slapd? I don't know Windows programming at all.

Yes, it's relevant. People running on Windows should probably boot with /usepmtimer to make sure the ACPI timer is used (which runs at 3.5MHz). Then again, the simplest solution is "don't run mission-critical servers on Windows" because the platform is so completely inadequate, for this and many other reasons. Probably should read this as well http://support.microsoft.com/kb/895980

Of course, not all of this uncertainty is Microsoft's fault - AMD documented that their dual-core processors would keep their TSCs in sync between both cores, but in reality the TSCs never stay in sync. So if you happen to be running an old-enough Windows release, written when the TSC was still believed to be a reliable clock source, you may have problems unless you explicitly tell Windows to use the ACPI PM timer.

If you're running on a very old motherboard that doesn't support ACPI, you won't have a PM timer to use; but in that case you're probably also running with a processor that doesn't have TSC issues.