Running more tests, seems a bit random either with acpihpet0 or tsc, you
are right. So please disregard, I have associated the wrong thing.
Happens either with debian or the alpine, and either timecounter.
Sometimes vm process maxes out, sometimes it's ok.
Thanks!
---
Oriol Demaria
2FFED630C16E4FF8
On 29/08/2019 22:53, Mike Larkin wrote:
On Thu, Aug 29, 2019 at 10:29:40PM +0100, Oriol Demaria wrote:
I have been following the patching of the TSC. So I don't have
problems on
the Ryzen now using TSC with the mouse and so on, but I have a problem
with
the vms. The CPU maxes out on the Ryzen 5 (on Intel is fine) when I
run a vm
(debian).
But when I change the timecounter back to acpihpet0 and restart vmd,
the CPU
usage is normal again. Does anyone else has this problem?
Regards.
--
Oriol Demaria
2FFED630C16E4FF8
40400 _vmd 10 0 8195M 505M idle fsleep 1:35 1.46% vmd
Not here. That's an ubuntu19 vm on Ryzen 7.
Does this happen all the time or is it random? We do have a
longstanding
well known bug where a VM can get stuck and spin to 100%, that's
unrelated
to TSC though.
-ml