Re: Thread Pre-Emption Question . . .

Tech-Archive recommends: Fix windows errors by optimizing your registry

+1

--
HLS

m wrote:
I'm glad that you have worked somthing out. If you still have problems, it might be better to start a new thread.


"Rob" <Rob@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote in message news:760257F7-CCC7-4CDA-92B7-81718333015A@xxxxxxxxxxxxxxxx
Hi All,

Thank you all for your replies, it appears I've sparked quite a debate!! I
think m hit the nail on the head early on, I'm trying to debug a situation
where the processing is taking too long and we're pretty much convinced the
issue doesn't lie in our code. We have a lot of preformance timing blocks in
our code (using QueryPerformaceTimer) so can see where it's taking to long
(it appears in a number of random sections of code for all the samples we've
seen) but the sections of code in question shouldn't take very long, which is
why be believe the thread is getting de-scheduled. We'd like to be able to
prove it but it appears that won't be so easy, I'll have a bit more of a play
with GetThreadTimes.

Regarding the I/O, all out network data writting & reading does happen on
different threads using a protected messages queuing mechanism, which is
initated from the main processing thread. My comment that I think spured
this debate was just that I didn't want to increase the base priority of our
process too high so that the OS networking services would get enought
processing time to do their thing.

I won't post the message queueing code here though ;)

Cheers
Rob.


.

Relevant Pages

  • Re: Thread Pre-Emption Question . . .
    ... I'm glad that you have worked somthing out. ... Thank you all for your replies, it appears I've sparked quite a debate!! ... all out network data writting & reading does happen on ... different threads using a protected messages queuing mechanism, ...
    (microsoft.public.win32.programmer.kernel)
  • Re: Thread Pre-Emption Question . . .
    ... I think m hit the nail on the head early on, I'm trying to debug a situation where the processing is taking too long and we're pretty much convinced the issue doesn't lie in our code. ... Regarding the I/O, all out network data writting & reading does happen on different threads using a protected messages queuing mechanism, which is initated from the main processing thread. ... My comment that I think spured this debate was just that I didn't want to increase the base priority of our process too high so that the OS networking services would get enought processing time to do their thing. ...
    (microsoft.public.win32.programmer.kernel)
  • Re: Creating my own "identity" type
    ... The replies weren't really a debate about keys but a response to your ... Since you didn't specify ...
    (microsoft.public.sqlserver.programming)
  • Re: Whats with all the punnery?
    ... r norman wrote: ... a serious debate. ... The replies are likewise just as quick ...
    (talk.origins)
  • Re: Religion is "the new social evil"
    ... }> Which part of my previous replies to you did you not understand, ... }> to debate with at least a suspicion of good faith. ... Seems rather a head-in-sand attitude from the Christians ... 7/7 etc are reported as religious fundamentalism. ...
    (uk.religion.christian)