Why would it be simpler or higher performance? Seems like it would be both more complicated and lower performance?
-
This Tweet is unavailable.
-
-
This Tweet is unavailable.
-
Replying to @Jonathan_Blow
I guess I just don't see how that paranoia comes into play here. When you have large numbers of cores, like you are suggesting, then you typically just aren't very oversubscribed. So you shouldn't have to worry about preemption in the middle of locks anyway?
1 reply 0 retweets 1 like -
This Tweet is unavailable.
-
This Tweet is unavailable.
-
Replying to @Jonathan_Blow
Yes, I agree with the approach completely. I just see preemptive multitasking as the _simpler_ option, because all it is is a timer interrupt you add on top of the existing (as-cooperative-as-it-wants-to-be) code that takes care of any unusual situations.
1 reply 0 retweets 2 likes
Unlike garbage collection, which must be run as a matter of course, preemption interrupts can be scheduled as infrequently as you would like, and even not used at all on certain cores if you know you want them dedicated. It is a nicely flexible model, whereas cooperative isn't.
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.