4205Re: a fix for the multi-core bug in the windows build
- Apr 11, 2012Steve and Bill, Thanks for the feedback. To clarify, this is what I see when using the stock build of OPML editor(v10.1a9):
and crashes on shutdown:
Running a hard loop script spreads the load across all cores:
Steve, Since you're seeing isolation on a single core that would explain your stability.
Bill, No other tests. Just wanted to make sure it was isolating properly on more than just my box. I think it's definitely architecture related. Steve's experience seems to bear this out, since I've seen this same behaviour on XP with dual-core AMD chips, but he's not seeing it on his quad-core machine. The specific architecture + OS combination that triggers it is a mystery to me.
--- In firstname.lastname@example.org, Bill Kearney <wkearney@...> wrote:
> I used it on core 2 (1-8 of an i7) running on a Windows 2008 server.
> Nails the core at 100%. Are there other tests you'd like to see run on
> it? Anything else that would trigger problems? Something predictable
> and repeatable would be good.
> Does it present problems on just multi-core single chip processors, on
> multiple processors, or both? Is this a core issue or processor issue?
> Is it architecture related? As in SMP vs NUMA?
> And be sure the processor affinity can be set somewhere externally. It
> would be bad form to hard code it to a given core number.
> On 4/10/2012 3:18 PM, sect62 wrote:
> > I've been hammering at this one for a while with no luck. This fix looks promising though:
> > http://worknotes.southernbread.org/bugs/4101285051AmByDj
> > I'd love to hear some feedback.
- << Previous post in topic Next post in topic >>