Re: Questions about NGPT vs NPTL
Dan Kegel <[email protected]>
| Newsgroups | gmane.linux.ngpt.devel |
|---|---|
| Message-ID | <[email protected]> |
[email protected] wrote: > If NGPT is faster than NPTL why you decided to stop the development? > Just because NPTL is already incorporated in glibc? NGPT may or may not be faster then NPTL; the differences are probably minor. And since the maintainer of glibc is dead-set on NPTL, the chances of overriding him are slim. > It is possible that you can contribute to NPTL to achieve the > same (or better) performance than NGPT has today? Same, sure. Better, not so sure. > What is the best model, 1:1 or M:N? That remains to be seen, and may depend on whether you're running a single processor or a multiCPU machine. However, note that Sun is dropping M:N in favor of 1:1, and FreeBSD wants to do M:N, but had to focus on 1:1 to get anything out the door. If those two projects' experience is a guide, then M:N appears to be a lot of work for not much benefit. See also http://kegel.com/c10k.html#threaded > You are stopped or are you active at this time? Are you working on > threads today or the project was cancelled? I'm not an NGPT team member, so take this with a grain of salt: I believe NGPT is in maintenance mode; bugfixes will be released, so existing projects using NGPT can keep using it for the time being without too much fear. However, new projects should use NGPT. - Dan -- Dan Kegel http://www.kegel.com http://counter.li.org/cgi-bin/runscript/display-person.cgi?user=78045