[Interest] Program runs only on four core processors.... sometimes.

Guido Seifert wargand at gmx.de
Fri Mar 22 20:33:28 CET 2013


Hi, me again. :-)

Not sure about priority inversion. When I check the network with netstat -an or the Resource Monitor the program seems to run. The userinterface, which connects also via tcp to the main program does not seem to come through. When I ask uncle Wikipedia, it looks to me more like resource starvation.

Yep, just seen while I wrote this mail. The program is not frozen... erm... sort of. What takes about 2-3 seconds with four corse seems to take ~15 minutes with one core. Never waited that long before. :-(

Not the result I would like, but at least consistent with what I would expect: Too slow to be usable, but at least running. I just did not see where a real deadlock could have come from.

Guido

On Wed, 20 Mar 2013 16:07:10 -0700
Thiago Macieira <thiago.macieira at intel.com> wrote:

> On quarta-feira, 20 de março de 2013 23.38.56, Guido Seifert wrote:
> > > Since decreasing the number of cores makes it work, 
> > 
> > Nope, increasing makes it work. But apparently not always. Or not for all
> > machines. Not sure, which of the two scenarios you described I should
> > prefer now. Even if I had the choice.
> 
> Sorry. I had written "Since increasing [...] makes to work", then I went back 
> to rewrite it. I ended up making it sound opposite of what I wanted.



More information about the Interest mailing list