> I've been recently working on a agent-based simulation system which is > intended to run with massive numbers of threads in the 15,000-30,000 > range. Thanks for the data points! You've probably considered this already, but switching to a state-machine model with fewer (1?) kernel-level threads may give you massive speed gains. Most kernels (especially Linux) perform very poorly when the # of active threads exceeds the # of CPUs by a large amount. (in fact, if you are on Linux, I also encourage you to try out some of the patches people have developed to increase the scalability of the scheduler - currently it is O(n) in the # of runnable proceses - check the Linux Kernel Mailing List archives...) Regards, Dan
RetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4