I have been experimenting with mixing v8 and user-mode scheduling. I have found it can be faster than node, when both have an affinity set for 1 processor. I have hopes in getting the speed up enough to have user-mode scheduling as an alternative to spinning up real threads as node does now. >From what I've read, user-mode threads switch much faster than real threads. Though they are not nearly as cheap as erlang's lightweight processes.
On Tue, Aug 13, 2013 at 5:54 AM, Kevin Swiber <[email protected]> wrote: > On Tue, Aug 13, 2013 at 12:08 AM, Tristan Slominski < > [email protected]> wrote: > >> >> >>> With the design I'm proposing, creating an actor will spawn a new >>> thread with its own event loop waiting to receive messages. I'm not sure >>> on the cost of that, but it would be interesting to find out. >>> >> >> What about a go-routine-like model (instead of thread spawning)? >> > > I spoke to a colleague about this a couple of days ago. It's worth > investigating, but will likely take much longer to implement (unless an > expert in this is willing to help). The thought behind using something > like Intel's Threading Building Blocks is that I don't have to implement > the task and scheduler abstractions and can hopefully build a proof of > concept in less time. This is more of a constraint in my own scope of > knowledge and experience. I try to learn Everything, but fortunately for > the Universe, there is more knowledge to be learned than time remaining in > my life. Scaling out with volunteers looking to contribute is certainly on > the table. > > >> >> Sure, but those addresses don't work when you scale beyond a single >>> process. I'm not sure how to solve that problem yet. >>> >> >> Yes, Mark's references seem a step in the right direction. Ken uses >> unguessable web-keys (referenced in the paper he mentioned: >> http://eros-os.org/pipermail/cap-talk/2012-September/015386.html). >> >>> This thread is turning into an ACM binder. I love it. Thanks! > > Cheers, > > -- > Kevin Swiber > Projects: https://github.com/kevinswiber > Twitter: @kevinswiber > > -- > -- > Job Board: http://jobs.nodejs.org/ > Posting guidelines: > https://github.com/joyent/node/wiki/Mailing-List-Posting-Guidelines > You received this message because you are subscribed to the Google > Groups "nodejs" group. > To post to this group, send email to [email protected] > To unsubscribe from this group, send email to > [email protected] > For more options, visit this group at > http://groups.google.com/group/nodejs?hl=en?hl=en > > --- > You received this message because you are subscribed to the Google Groups > "nodejs" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > For more options, visit https://groups.google.com/groups/opt_out. > > > -- -- Job Board: http://jobs.nodejs.org/ Posting guidelines: https://github.com/joyent/node/wiki/Mailing-List-Posting-Guidelines You received this message because you are subscribed to the Google Groups "nodejs" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/nodejs?hl=en?hl=en --- You received this message because you are subscribed to the Google Groups "nodejs" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/groups/opt_out.
