Make sure you set X10_NTHREADS. See http://x10.codehaus.org/For+Users, specifically the "X10 Environment Variables" section.
See the FAQ -- How do I control the size of the thread pool in a place. http://x10.codehaus.org/FAQ#FAQ-HowdoIcontrolthesizeofthethreadpoolinaplace%3F On 11/3/2010 5:11 PM, Russel Winder wrote: > I am fairly confident I am just missing something very simple but . . . > > I have a (trivial) embarrassingly parallel scatter/gather code which > compiles fine in both Java back-end and C++ back-end modes. The Java > backend-version runs nicely using all the cores it can for the number of > places specified in the code -- though the code runs excruciatingly > slowly compared to Java or Scala. The point is I see the scaling that > should be present. Using the C++ backend, it is nice and fast > (essentially as fast as C++ coded version using asynchronous function > calls a la C++0x) but it only ever runs on one core. So how do I get > the runx10 command to map places to cores? > > Am I just being very stupid and missing the line that tells me how to do > this? > > Thanks. > > PS I remembers to switch off PGP signing before sending this one so > hopefully it doesn't get rejected. ------------------------------------------------------------------------------ The Next 800 Companies to Lead America's Growth: New Video Whitepaper David G. Thomson, author of the best-selling book "Blueprint to a Billion" shares his insights and actions to help propel your business during the next growth cycle. Listen Now! http://p.sf.net/sfu/SAP-dev2dev _______________________________________________ X10-users mailing list X10-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/x10-users
