Cool. I have changed my code towards using ranges, as well. Also, my
map only emits prime numbers and no none-prime numbers. This is for
better memory requirements. When I run ranges of 1,000,000 numbers
wide, I can see that my 4 cores are 100% busy and kernel time is close
to nothing. Though, the processor utilization seems fine.
When do you think you'll fix the problem with iterating over the results?
Regards,
Christian
On Sat, Aug 22, 2009 at 3:07 PM, Craig
Henderson
I'm afraid to admit that I f***ed up. The program is now working. I invalidated the iterators with a for loop before the for_each call. When I remove my for loop, the results are inside the output files.
Cool stuff, Christian
Great! I've added my version to the sandbox at http://svn.boost.org/svn/boost/sandbox/libs/mapreduce/examples/prime/
This includes a modified number_source that provides a range for each map task to work on, reducing locking contention.
Regards -- Craig
_______________________________________________ Boost-users mailing list Boost-users@lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users