On Fri, Nov 15, 2013 at 12:35:22AM +0400, Alexander Pakhomov wrote:
dd unenc sp:
sys 17% int 9%
164 MB/s
dd_unenc_mp
30 sys 50 int
200 MB/s
this roughly shows that a lot of CPUs time is wasted in interrupts
in the MP case, probably spinning to grab the kernel_lock held by
the other CPU.
Hi.
I discovered high CPU0 interruptions load (amd64) in various cases.
wifi (athn, urtwn): up to 100% at 2 MB/s download. Usually 50%. Grows faster
then linear with net load. Sometimes system freezes for about a minute.
Web speed tests consume significantly less CPU then wget/firefox/ktorrent
On Thu, Nov 14, 2013 at 01:16:41PM +0400, Alexander Pakhomov wrote:
Hi.
I discovered high CPU0 interruptions load (amd64) in various cases.
wifi (athn, urtwn): up to 100% at 2 MB/s download. Usually 50%. Grows faster
then linear with net load. Sometimes system freezes for about a minute.
Web
On 11/14/2013 03:19 PM, Alexandre Ratchov wrote:
On Thu, Nov 14, 2013 at 01:16:41PM +0400, Alexander Pakhomov wrote:
1) What interruptions load should be considered normal? I used to
think even 10% is too high.
yes, though it depends on the workload
I want to investigate the problem
dmesg?
14.11.2013, 17:20, Alexandre Ratchov a...@caoua.org:
On Thu, Nov 14, 2013 at 01:16:41PM +0400, Alexander Pakhomov wrote:
Hi.
I discovered high CPU0 interruptions load (amd64) in various cases.
wifi (athn, urtwn): up to 100% at 2 MB/s download. Usually 50%. Grows faster
then linear
+0400, Alexander Pakhomov wrote:
Hi.
I discovered high CPU0 interruptions load (amd64) in various cases.
wifi (athn, urtwn): up to 100% at 2 MB/s download. Usually 50%. Grows
faster
then linear with net load. Sometimes system freezes for about a minute.
Web speed tests consume
I was wrong. Kernel panics with splassert. So bug report I just sent is
somehow connected with interruptions.
15.11.2013, 00:44, Alexander Pakhomov ker0...@yandex.ru:
By the way boot sync in ddb cause kernel panic (both mp and sp). Will send
next bug report in bugs@. Maybe this is connected
8 matches
Mail list logo