worker thread performance question

[Julian Elischer]

Petri Helenius wrote:

> Julian Elischer wrote:
>
>>
>>
>> Petri Helenius wrote:
>>
>>>
>>> With libpthread is it usually optimal to have as many worker threads 
>>> (CPU bound stuff) as kern.threads.virtual_cpu or have, say double 
>>> the number so that there is always a thread in the run queue when 
>>> another hits a mutex or sleep? 
>>
>>
>>
>> Are they always runnable?
>
>
> Almost, except when they run into shared structures which obviously 
> are minimized by design.
>
>> theoretically you can schedule as many as you want.
>> any number > NCPU should keep teh system busy, but
>> I'm not sure I fully understand the question.
>
>
> The question was aimed towards if larger runqueue takes more CPU to 
> maintain than a shorter one does and if threads are involuntarily 
> switched. 


yes and yes.

>
>
> Pete
>
>>
>>>
>>>
>>> Pete
>>>
>>> _______________________________________________
>>> freebsd-threads at freebsd.org mailing list
>>> http://lists.freebsd.org/mailman/listinfo/freebsd-threads
>>> To unsubscribe, send any mail to 
>>> "freebsd-threads-unsubscribe at freebsd.org"
>>
>>
>>
>>