[R-sig-hpc] foreach + doMC not fully parallel?
Brian D Peyser PhD
bpeyser at jhmi.edu
Mon Aug 30 18:23:08 CEST 2010
On Sun, 2010-08-29 at 11:56 +0800, Xie Chao wrote:
> I suggest you try:
> 1) run your R code from R directly
> 2) test similar code without IO from hard disk
> 3) If combining 1) and 2) still give you the similar result, could you
> provide a simple reproducible code? Especially the part you mentioned
> as "lots of computation".
> Xie Chao
1 & 2 I had already tried.
I did a reboot and %dopar% started working fully again. I then did
suspend/resume and the problem came back. i rarely reboot and use
suspend/resume all the time (running on my laptop). Most likely it's a
kernel problem. I will have to investigate a little more thoroughly, but
it's obvious now that this is not an R problem.
Thanks everyone here who gave advice! Seems time to ask the kernel folks
at Ubuntu. Definitely there's a bug here (maybe something to do with
SCI_EN--I'll have to troubleshoot).
More information about the R-sig-hpc