>> I propose that the developer can choose that new preemptive >> threads will be executed in the currently running process. >> > There's no choice, they have to be because of data sharing. > The only option is whether context switches are pre-emptive > or not -- however the two things are semantically distinct. > > Pre-emption requires mutexing data accesses, non-premption > typically doesn't. This means you could downgrade a pthread > to an fthread without changing the semantics (provided of > course it doesn't block). > > I agree some way of 'hinting' at this is needed, but > it isn't clear how to do that... think again of the > recursive quadrature routine in the SR language, > which spawns 2^n threads .. the hints can't be static > in that case. > Can the software "Automatic Parallelization Environment for Network of Workstations" help here to achieve useful concurrency and efficiency? http://www.cs.wayne.edu/~haj/html/ape4now.html
How do you think about to cooperate with distributed and shared memory code? http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2006/n2044.html Regards, Markus ------------------------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 _______________________________________________ Felix-language mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/felix-language
