We’re bringing a new mid-tier compiler to Chrome. Maglev is a just-in-time compiler that can quickly generate performant machine code for all relevant functions within the first one-hundredth of a second. It reduces overall CPU time to compile code while also saving battery life. Our measurements show Maglev has provided a 7.5 percent improvement on Jetstream and a 5 percent improvement in Speedometer. Maglev will start rolling out in Chrome version 114, which begins release on June 5.
Let’s hope making benchmarks run faster also makes actual websites load faster.
Greetings Thom.
> Let’s hope making benchmarks run faster also makes actual websites load faster.
Despite publishing great content, sometimes I feel like you have an overly negative bias to certain technologies like Google or BitCoin. That is your right of course and criticism was not a bad thing when you’d also link it to suggestion how to actually improve the situation.
For the example above, this compiler is likely for the JavaScript code and it will simply compile that JavaScript code faster draining less power. It won’t make your WebSite loading any faster. And this not Googles fault at all — we all know who pushed for JavaScript when a strongly typed, sandbox with proper GC was available.
I don’t use Chrome and I don’t like JavaScript and this development won’t load websites faster. But I am happy that they are working on it as it will provide options and options are never wrong or bad.
Cheers and keep up the great work!
These changes are supposed to make a difference for end users experience of using websites, not just battery life. Read the descriptions of the benchmarks and what they are trying to measure.
Bitcoin should be destroyed. The world would be better off if it and any of the other cryptos didn’t exist. That’0s not opinion but a very obvious fact supported by a wealth of data.
As for anti Google bias? Eh, all tech companies should be criticized, and any regular reader should know that there really aren’t any prefect trillion dollar tech companies. They are not your friends and not everything they touch turns to gold.
“within the first one-hundredth of a second.”
On a 10MHz PIC-micorcontroller?
WOW!
That’s impressive!
…
What a stupid metric…
smarhIt,
You’re right this stuck out to me as a bad metric when I read it as well. The time it takes to run an arbitrary benchmark on arbitrary hardware isn’t meaningful to users. Still, optimizations that improve experience and battery life are good, that’s the important thing 🙂
Maglev (Hebrew) – מגלב – whip
Not that I’m paranoid …
You know what a maglev train is right?
Between 5 and 7.5 percent improvement doesn’t sound a whole lot, and it’s just the JIT, the JITed code won’t run any faster. And no doubt this has cost millions of dollars to develop.
I don’t know why people think this is only improving battery life or compiled time. It literally says it improves the speed of the JIT code, according to the benchmarks. Or am I missing something obvious?