Re: xeon 2.8GHz SMP/NOT test results
"Attilio Rao" <[email protected]>
| Newsgroups | gmane.os.freebsd.devel.smp |
|---|---|
| Message-ID | <[email protected]> |
2006/10/16, Kian Mohageri <[email protected]>: > On 10/16/06, Ivan Voras <[email protected]> wrote: > > > > Ok, but take care so that the machine is *completely* idle while you > > doing the tests, otherwise they don't mean anything. > > > > > Understood. > > I noticed somebody suggested trying SMP with hyperthreading disabled in the > BIOS, and I'm a little bit confused as to what sort of results should I > expect from that. Could anybody enlighten me about how the OS will see/make > use of this? > > My next test(s) will probably be: > > - "SMP w/ HTT" vs. "GENERIC w/ HTT disabled in BIOS" vs. "SMP w/ HTT > disabled in BIOS" > - both lmsensors and unixbench tests > - timecounter comparisons I think that you might do a real comparison using the same kernel w/ only these slight differences (read: not using GENERIC vs. custom-kernel): disable HTT in BIOS: - kernel w/o SMP (w/o PREEMPTION+FULL_PREEMPTION) - kernel w/ SMP (w/o PREEMPTION+FULL_PREEMPTION) - kernel w/o SMP (w/ PREEMPTION+FULL_PREEMPTION) - kernel w/ SMP (w/ PREEMPTION+FULL_PREEMPTION) enable HTT in BIOS, active HTT in the kernel and do again these tests (you might have an 8 cases situation to show). Attilio -- Peace can only be achieved by understanding - A. Einstein _______________________________________________ [email protected] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-smp To unsubscribe, send any mail to "[email protected]"