node
node copied to clipboard
Published
20 hours ago •
nodejs
nodejs
console: make console.log resilient to built-in prototypes tempering
Benchmark CI: https://ci.nodejs.org/view/Node.js%20benchmark/job/benchmark-node-micro-benchmarks/944/
Do we really need a test for this?
I added the test because it causes significant perf regressions to the events and streams benchmarks:, and if we are going to make that tradeoff, we should have a test otherwise it might be reverted in the future.
confidence improvement accuracy (*) (**) (***)
events/ee-emit.jslisteners=5 argc=0 n=2000000 * 3.38 % ±3.05% ±4.07% ±5.31%
events/ee-listeners.jsraw='false' listeners=50 n=5000000 ** 2.15 % ±1.54% ±2.04% ±2.66%
events/ee-once.jsargc=0 n=20000000 *** -10.08 % ±0.88% ±1.17% ±1.52%
events/ee-once.jsargc=1 n=20000000 *** -6.86 % ±2.52% ±3.39% ±4.50%
events/ee-once.jsargc=4 n=20000000 *** -8.15 % ±0.99% ±1.32% ±1.72%
events/ee-once.jsargc=5 n=20000000 *** -8.20 % ±1.36% ±1.82% ±2.40%
streams/creation.jskind='transform' n=50000000 *** -5.39 % ±2.97% ±3.96% ±5.17%
streams/creation.jskind='writable' n=50000000 *** -8.28 % ±1.46% ±1.95% ±2.55%
streams/readable-boundaryread.jstype='string' n=2000 * 4.07 % ±3.98% ±5.30% ±6.90%
confidence improvement accuracy (*) (**) (***)
events/ee-add-remove.jsn=1000000 -2.58 % ±3.38% ±4.52% ±5.93%
events/ee-emit.jslisteners=10 argc=0 n=2000000 -1.35 % ±2.12% ±2.85% ±3.75%
events/ee-emit.jslisteners=10 argc=10 n=2000000 0.18 % ±0.73% ±0.97% ±1.26%
events/ee-emit.jslisteners=10 argc=2 n=2000000 0.88 % ±1.30% ±1.73% ±2.25%
events/ee-emit.jslisteners=10 argc=4 n=2000000 0.20 % ±0.96% ±1.28% ±1.66%
events/ee-emit.jslisteners=1 argc=0 n=2000000 0.52 % ±3.82% ±5.10% ±6.65%
events/ee-emit.jslisteners=1 argc=10 n=2000000 1.31 % ±3.56% ±4.74% ±6.18%
events/ee-emit.jslisteners=1 argc=2 n=2000000 2.08 % ±4.79% ±6.38% ±8.32%
events/ee-emit.jslisteners=1 argc=4 n=2000000 0.29 % ±4.39% ±5.84% ±7.60%
events/ee-emit.jslisteners=5 argc=0 n=2000000 * 3.38 % ±3.05% ±4.07% ±5.31%
events/ee-emit.jslisteners=5 argc=10 n=2000000 -0.32 % ±2.04% ±2.72% ±3.55%
events/ee-emit.jslisteners=5 argc=2 n=2000000 -0.86 % ±1.72% ±2.28% ±2.97%
events/ee-emit.jslisteners=5 argc=4 n=2000000 0.60 % ±1.50% ±2.00% ±2.61%
events/ee-listener-count-on-prototype.jsn=50000000 -1.42 % ±3.29% ±4.42% ±5.82%
events/ee-listeners.jsraw='false' listeners=50 n=5000000 ** 2.15 % ±1.54% ±2.04% ±2.66%
events/ee-listeners.jsraw='false' listeners=5 n=5000000 2.24 % ±5.27% ±7.08% ±9.35%
events/ee-listeners.jsraw='true' listeners=50 n=5000000 -0.90 % ±1.67% ±2.23% ±2.90%
events/ee-listeners.jsraw='true' listeners=5 n=5000000 -0.41 % ±2.95% ±3.92% ±5.11%
events/ee-once.jsargc=0 n=20000000 *** -10.08 % ±0.88% ±1.17% ±1.52%
events/ee-once.jsargc=1 n=20000000 *** -6.86 % ±2.52% ±3.39% ±4.50%
events/ee-once.jsargc=4 n=20000000 *** -8.15 % ±0.99% ±1.32% ±1.72%
events/ee-once.jsargc=5 n=20000000 *** -8.20 % ±1.36% ±1.82% ±2.40%
events/eventtarget.jslisteners=10 n=1000000 1.40 % ±5.79% ±7.72% ±10.07%
events/eventtarget.jslisteners=1 n=1000000 -1.16 % ±3.42% ±4.58% ±6.00%
events/eventtarget.jslisteners=5 n=1000000 2.71 % ±5.36% ±7.20% ±9.51%
streams/creation.jskind='duplex' n=50000000 -2.42 % ±2.82% ±3.75% ±4.89%
streams/creation.jskind='readable' n=50000000 -2.30 % ±2.60% ±3.46% ±4.50%
streams/creation.jskind='transform' n=50000000 *** -5.39 % ±2.97% ±3.96% ±5.17%
streams/creation.jskind='writable' n=50000000 *** -8.28 % ±1.46% ±1.95% ±2.55%
streams/pipe.jsn=5000000 1.84 % ±2.65% ±3.54% ±4.63%
streams/pipe-object-mode.jsn=5000000 -0.61 % ±4.02% ±5.36% ±7.00%
streams/readable-async-iterator.jssync='no' n=100000 -0.65 % ±2.18% ±2.90% ±3.79%
streams/readable-async-iterator.jssync='yes' n=100000 -1.40 % ±4.42% ±5.88% ±7.66%
streams/readable-bigread.jsn=1000 2.75 % ±5.48% ±7.29% ±9.49%
streams/readable-bigunevenread.jsn=1000 -2.08 % ±5.99% ±7.98% ±10.40%
streams/readable-boundaryread.jstype='buffer' n=2000 1.45 % ±2.81% ±3.74% ±4.87%
streams/readable-boundaryread.jstype='string' n=2000 * 4.07 % ±3.98% ±5.30% ±6.90%
streams/readable-readall.jsn=5000 0.95 % ±14.35% ±19.09% ±24.85%
streams/readable-unevenread.jsn=1000 2.20 % ±3.63% ±4.84% ±6.31%
streams/writable-manywrites.jslen=1024 callback='no' writev='no' sync='no' n=2000000 1.02 % ±1.78% ±2.37% ±3.09%
streams/writable-manywrites.jslen=1024 callback='no' writev='no' sync='yes' n=2000000 -5.32 % ±6.95% ±9.33% ±12.30%
streams/writable-manywrites.jslen=1024 callback='no' writev='yes' sync='no' n=2000000 -1.82 % ±2.61% ±3.47% ±4.52%
streams/writable-manywrites.jslen=1024 callback='no' writev='yes' sync='yes' n=2000000 -2.80 % ±3.67% ±4.88% ±6.36%
streams/writable-manywrites.jslen=1024 callback='yes' writev='no' sync='no' n=2000000 0.68 % ±1.11% ±1.48% ±1.92%
streams/writable-manywrites.jslen=1024 callback='yes' writev='no' sync='yes' n=2000000 0.13 % ±3.96% ±5.27% ±6.86%
streams/writable-manywrites.jslen=1024 callback='yes' writev='yes' sync='no' n=2000000 -1.83 % ±2.45% ±3.26% ±4.25%
streams/writable-manywrites.jslen=1024 callback='yes' writev='yes' sync='yes' n=2000000 -1.68 % ±3.32% ±4.42% ±5.75%
streams/writable-manywrites.jslen=32768 callback='no' writev='no' sync='no' n=2000000 1.23 % ±3.97% ±5.34% ±7.08%
streams/writable-manywrites.jslen=32768 callback='no' writev='no' sync='yes' n=2000000 -2.05 % ±3.50% ±4.70% ±6.22%
streams/writable-manywrites.jslen=32768 callback='no' writev='yes' sync='no' n=2000000 -2.54 % ±2.61% ±3.50% ±4.60%
streams/writable-manywrites.jslen=32768 callback='no' writev='yes' sync='yes' n=2000000 -0.14 % ±1.74% ±2.32% ±3.02%
streams/writable-manywrites.jslen=32768 callback='yes' writev='no' sync='no' n=2000000 -0.37 % ±1.65% ±2.20% ±2.88%
streams/writable-manywrites.jslen=32768 callback='yes' writev='no' sync='yes' n=2000000 -1.69 % ±3.43% ±4.57% ±5.95%
streams/writable-manywrites.jslen=32768 callback='yes' writev='yes' sync='no' n=2000000 -1.25 % ±1.98% ±2.63% ±3.43%
streams/writable-manywrites.jslen=32768 callback='yes' writev='yes' sync='yes' n=2000000 0.39 % ±2.02% ±2.69% ±3.51%
Be aware that when doing many comparisons the risk of a false-positive
result increases. In this case, there are 55 comparisons, you can thus
expect the following amount of false-positive results:
2.75 false positives, when considering a 5% risk acceptance (*, **, ***),
0.55 false positives, when considering a 1% risk acceptance (**, ***),
0.06 false positives, when considering a 0.1% risk acceptance (***)
V8 9.0 update didn't fix the perf regression:
events/ee-once.jsargc=0 n=20000000 *** -12.11 % ±2.77% ±3.73% ±4.95%
events/ee-once.jsargc=1 n=20000000 *** -8.18 % ±1.67% ±2.22% ±2.89%
events/ee-once.jsargc=4 n=20000000 *** -8.07 % ±2.53% ±3.37% ±4.39%
events/ee-once.jsargc=5 n=20000000 *** -9.74 % ±1.74% ±2.32% ±3.04%
https://ci.nodejs.org/view/Node.js%20benchmark/job/benchmark-node-micro-benchmarks/962/