I see massive changes when I increase THROTTLE, and the whole thing times out when I use THROTTLE=32, which is what I expect. Run it with the latest version of Puppeteer (5.3.0) using THROTTLE=4 node longtask.js (the THROTTLE variable controls the CPU throttle rate). You'll get almost the exact same numbers.Ĭonst puppeteer = require ( 'puppeteer' ) ( async function ( ) ) ( ) Run that, then run it again without the -chrome.CPUThrottlingRate 32 part and compare the output. bin/browsertime.js -cpu -video=false -visualMetrics=false -n 5 -chrome.CPUThrottlingRate 32
I thought this might be Docker-related, but running browsertime directly gives me the same results. The logs claim it's using CPU throttling, but the total blocking time is way too low (1.78s), and when I run into it without throttling, I get almost the exact same numbers. INFO: HTML stored in /sitespeed.io/sitespeed-result//-00-30-33 INFO: Start collecting Lighthouse result for iteration 1 INFO: Got analysed from Google Page Speed Insights INFO: Use Chrome trace categories: -*,disabled-by-default-lighthouse,v8,v8.execute,er_timing,devtools.timeline,disabled-by-default-devtools.timeline, INFO: Running tests using Chrome - 5 iteration(s)
INFO: Sending url to test on Page Speed Insights INFO: Will run Lighthouse tests after Browsertime has finished