Re: [ANN] core.async 0.2.382

2016-06-23 Thread Alan Moore
Shouldn't the default be a value related to the number of cores like the original too large default was? Maybe not exactly (.availableProcessors (Runtime/getRuntime)), as suggested by Fluid Dynamics, but something similar? In just about every case where the # cores != 8 you will be under/over p

Re: [ANN] core.async 0.2.382

2016-06-23 Thread Alex Miller
First, note that this is a *max* pool size. There is no "right" number. 42+2n was "a lot" (and sometimes too many). 8 is "less" and a reasonable number for 2-8 core machines common in servers. More important is that it's now configurable to a number that's good for you. On Wednesday, June 22, 2

Re: [ANN] core.async 0.2.382

2016-06-23 Thread Ning Sun
I guess this change is to encourage user to configure their own pool size based on their use case, instead of completely relying on the default. On 06/23/2016 11:33 AM, Daniel Compton wrote: > I knew I must be missing something. You can also set the core.async > threadpool size with > > (System/s

Re: [ANN] core.async 0.2.382

2016-06-22 Thread Daniel Compton
I knew I must be missing something. You can also set the core.async threadpool size with (System/setProperty "clojure.core.async.pool-size" "42") as long as that runs before any core.async code tries to use the threadpool. Some context on why the change was made to 8 would be good, but it is easy

Re: [ANN] core.async 0.2.382

2016-06-22 Thread Daniel Compton
Bumping this too. What was the context for this change? I don’t see any JIRA tickets linked in the commits with a rationale for why 8 was picked as the thread pool siz

Re: [ANN] core.async 0.2.382

2016-06-13 Thread Fluid Dynamics
On Monday, June 13, 2016 at 4:14:25 PM UTC-4, Alex Miller wrote: > > core.async 0.2.382 is now available. > > Try it via: [org.clojure/core.async "0.2.382"] > > 0.2.382 includes the following changes: > > - Change default dispatch thread pool max size to 8. > - Add Java system property clojure.cor