stream: improve Readable.read() performance

read() performance is improved most by switching from an array to a linked list for storing buffered data. However, other changes that also contribute include: making some hot functions inlinable, faster read() argument checking, and misc code rearrangement to avoid unnecessary code execution. PR-URL: #7077 Reviewed-By: Calvin Metcalf <[email protected]> Reviewed-By: Matteo Collina <[email protected]>

Author: mscdex

benchmark/streams/readable-bigread.js

@@ -0,0 +1,32 @@
+'use strict';
+
+constcommon=require('../common');
+constv8=require('v8');
+constReadable=require('stream').Readable;
+
+constbench=common.createBenchmark(main,{
+n: [100e1]
+});
+
+functionmain(conf){
+constn=+conf.n;
+constb=newBuffer(32);
+consts=newReadable();
+functionnoop(){}
+s._read=noop;
+
+// Force optimization before starting the benchmark
+s.push(b);
+v8.setFlagsFromString('--allow_natives_syntax');
+eval('%OptimizeFunctionOnNextCall(s.read)');
+s.push(b);
+while(s.read(128));
+
+bench.start();
+for(vark=0;k<n;++k){
+for(vari=0;i<1e4;++i)
+s.push(b);
+while(s.read(128));
+}
+bench.end(n);
+}

benchmark/streams/readable-bigunevenread.js

@@ -0,0 +1,32 @@
+'use strict';
+
+constcommon=require('../common');
+constv8=require('v8');
+constReadable=require('stream').Readable;
+
+constbench=common.createBenchmark(main,{
+n: [100e1]
+});
+
+functionmain(conf){
+constn=+conf.n;
+constb=newBuffer(32);
+consts=newReadable();
+functionnoop(){}
+s._read=noop;
+
+// Force optimization before starting the benchmark
+s.push(b);
+v8.setFlagsFromString('--allow_natives_syntax');
+eval('%OptimizeFunctionOnNextCall(s.read)');
+s.push(b);
+while(s.read(106));
+
+bench.start();
+for(vark=0;k<n;++k){
+for(vari=0;i<1e4;++i)
+s.push(b);
+while(s.read(106));
+}
+bench.end(n);
+}

benchmark/streams/readable-boundaryread.js

@@ -0,0 +1,33 @@
+'use strict';
+
+constcommon=require('../common');
+constv8=require('v8');
+constReadable=require('stream').Readable;
+
+constbench=common.createBenchmark(main,{
+n: [200e1]
+});
+
+functionmain(conf){
+constn=+conf.n;
+constb=newBuffer(32);
+consts=newReadable();
+functionnoop(){}
+s._read=noop;
+
+// Force optimization before starting the benchmark
+s.push(b);
+v8.setFlagsFromString('--allow_natives_syntax');
+eval('%OptimizeFunctionOnNextCall(s.push)');
+eval('%OptimizeFunctionOnNextCall(s.read)');
+s.push(b);
+while(s.read(32));
+
+bench.start();
+for(vark=0;k<n;++k){
+for(vari=0;i<1e4;++i)
+s.push(b);
+while(s.read(32));
+}
+bench.end(n);
+}

benchmark/streams/readable-readall.js

@@ -0,0 +1,32 @@
+'use strict';
+
+constcommon=require('../common');
+constv8=require('v8');
+constReadable=require('stream').Readable;
+
+constbench=common.createBenchmark(main,{
+n: [50e2]
+});
+
+functionmain(conf){
+constn=+conf.n;
+constb=newBuffer(32);
+consts=newReadable();
+functionnoop(){}
+s._read=noop;
+
+// Force optimization before starting the benchmark
+s.push(b);
+v8.setFlagsFromString('--allow_natives_syntax');
+eval('%OptimizeFunctionOnNextCall(s.read)');
+s.push(b);
+while(s.read());
+
+bench.start();
+for(vark=0;k<n;++k){
+for(vari=0;i<1e4;++i)
+s.push(b);
+while(s.read());
+}
+bench.end(n);
+}

benchmark/streams/readable-unevenread.js

@@ -0,0 +1,32 @@
+'use strict';
+
+constcommon=require('../common');
+constv8=require('v8');
+constReadable=require('stream').Readable;
+
+constbench=common.createBenchmark(main,{
+n: [100e1]
+});
+
+functionmain(conf){
+constn=+conf.n;
+constb=newBuffer(32);
+consts=newReadable();
+functionnoop(){}
+s._read=noop;
+
+// Force optimization before starting the benchmark
+s.push(b);
+v8.setFlagsFromString('--allow_natives_syntax');
+eval('%OptimizeFunctionOnNextCall(s.read)');
+s.push(b);
+while(s.read(12));
+
+bench.start();
+for(vark=0;k<n;++k){
+for(vari=0;i<1e4;++i)
+s.push(b);
+while(s.read(12));
+}
+bench.end(n);
+}