rxjs-cluster
v0.3.15
Published
Use Node.js cluster support to parallize Rx
Downloads
27
Readme
rxjs-cluster
Moore's law ceiling: "The biggest potential for improved performance is now, as I see it, on the software side [...] with parallelism which has so far not been [broadly] implemented."
(WIP: Working Beta)
Using Rx, maximize CPU usage in Node by using the new clusterMap that uses cluster/forked processes
var Rx = require('rx');
var Cluster = require('rxjs-cluster'); // import
var options = {};
var rc = new Cluster( options ); // instance
var Observable = Rx.Observable;
// Child function that returns raw value
function childTest(x) {
return "hello " + x + " from " + process.pid;
}
// Child function that returns an Observable
function childTest$(x) {
return Rx.Observable.range(0,3).map("hello " + x).toArray();
}
function master() {
Observable.from(['Jonathan', 'James', 'Edwin'])
.clusterMap('childTest')
.subscribe(
function(x) { console.log(x); },
function(x) { console.log('Err ' + x); },
function() { console.log('Completed'); }
);
Observable.from(['Jonathan', 'James', 'Edwin'])
.clusterMap('childTest$')
.subscribe(
function(x) { console.log(x); },
function(x) { console.log('Err ' + x); },
function() {
console.log('Completed');
rc.killall(); // kill all workers, clusterMap will no longer work
}
);
}
// Define number of workers, master entry point, worker functions
rc.entry(master, { 'childTest': childTest,
'childTest$': childTest$ });
// Or define leave the default number of workers to # of cpu cores
// rc.entry(master, { 'childTest': childTest, 'childTest$': childTest$ });