Nothing Special   »   [go: up one dir, main page]

Since the launch of Chrome in 2008, speed has been one of the 4 core principles that shape the work we do to deliver a highly performant browser. The V8 JavaScript compiler is a critical part of delivering maximum speed for the JavaScript that’s shipped on practically every web page. In our next post in The Fast and the Curious series, we are excited to share how improvements to the V8 engine are delivering up to 23% faster performance.


An important component of delivering a fast browser is fast JavaScript execution. In Chrome, that job is done by the V8 engine which executes over 78 years worth of JavaScript code on a daily basis. In M91 Chrome is now up to 23% faster with the launch of a new Sparkplug compiler and short builtin calls, saving over 17 years of our users' CPU time each day! Sparkplug is a new JavaScript compiler that fills the gap between needing to start executing quickly and optimizing the code for maximum performance. Short builtin calls optimize where in memory we put generated code to avoid indirect jumps when calling functions.



SparkplugThe V8 engine has multiple compilers which can make different tradeoffs throughout the various phases of executing JavaScript. Three years ago, we launched a new two-tier compiler system consisting of Ignition and Turbofan. Ignition is a bytecode interpreter whose job is to start executing the JavaScript with as little delay as possible. Turbofan is the optimizing compiler that generates high-performance machine code based on information gathered during JavaScript execution; as a result, it starts up more slowly than Ignition’s bytecode compiler. Sparkplug strikes a balance between Ignition and Turbofan in that it does generate native machine code but does not depend on information gathered while executing the JavaScript code. This lets it start executing quickly while still generating relatively fast code. For a complete technical deep dive into what it took to make this new engine, please see our V8 blog post.






Short builtins Short builtins is a mechanism by which the V8 engine optimizes the location in memory of generated code. When V8 generates CPU-specific code from JavaScript, it lays that code out in memory. This generated code will frequently call builtin functions, which are small snippets of code for handling common routines --everything from basic operations like adding two variables, to full-fledged functions in the JavaScript standard library. For some CPUs, calling functions that are further away from your generated code can cause CPU-internal optimizations (such as branch prediction logic) to fail. The fix for this is to copy the builtin functions into the same memory region as the generated code. This change is especially impactful for the new Apple M1 chip. Please see our V8 blog post to learn more about the impact across platforms of this feature.

Stay turned for many more performance improvements to come!

Posted by Thomas Nattestad, Chrome Product Manager


Data source for all statistics: Speedometer 2.0.
Labels: ,

Today, people have many ways to keep up with their favorite websites, including subscribing to mailing lists, notifications and RSS. It’s a lot for any one person to manage, so we’re exploring how to simplify the experience of getting the latest and greatest from your favorite sites directly in Chrome, building on the open RSS web standard. Our vision is to help people build a direct connection with their favorite publishers and creators on the web.

In the coming weeks, some Android users in the US on Chrome Canary may see an experimental Follow feature designed to help people get the latest content from sites they follow. Our goal for this feature is to allow people to follow the websites they care about, from the large publishers to the small neighborhood blogs, by tapping a Follow button in Chrome. When websites publish content, users can see updates from sites they have followed in a new Following section on the New Tab page:




Keeping a site’s RSS up-to-date will ensure Chrome can provide the latest content to users with this experiment. We will provide more guidance to web publishers as we learn and evaluate whether this feature will graduate from an experiment to a broader rollout in Chrome.

We welcome feedback from publishers, bloggers, creators, and citizens of the open web (like you!) on this experiment as we aim to build deeper engagement between users and web publishers in Chrome. You can also stay up-to-date and ask us questions via @GoogleCreators on Twitter or via email to webcreators@google.com. As part of this year’s Google I/O, we’ll be hosting a Meet Up for web publishers, creators and developers who would like to learn more, ask questions and share feedback. You can sign up for I/O (free this year) and register for the Following on the Open Web session, being held on May 19 (today) at 11 AM PT.

Posted by Janice Wong, Product Manager, Google Chrome
Labels:

Updates

A little over a year ago we announced our plans to reduce the granularity of information available from the User-Agent string, which is sent by default for every HTTP request. Shortly after, we made the decision to put this effort on pause so as not to create an additional migration burden on the web ecosystem in the early days of the COVID-19 pandemic. Since then, we’ve spent a lot of time gathering valuable feedback from the ecosystem, proposing ergonomic improvements to the User-Agent Client Hints API (UA-CH)—our proposed replacement for content negotiation and detection—as well as making web compatibility fixes.

UA-CH is now shipping by default in Chrome (since M89). We’ve also started the roll-out of both Client Hints Reliability mechanisms (Critical-CH & ACCEPT_CH) to address use cases where hints are needed on the first request. While we don’t yet have exact dates and milestones to announce for the planned User-Agent string reduction changes, we’re ready to resume our efforts on this front.

That said, we feel it's important to proceed in a way that gives the ecosystem and developers sufficient time to test use cases, provide feedback, and migrate to UA-CH where appropriate, which is why no User-Agent string changes will be coming to the stable channel of Chrome in 2021. Our intent with this post is to provide transparency into our thinking and roadmap early on so you can plan to adapt accordingly.

What is changing, and how?

We plan to gradually reduce, in a phased manner, the granularity of available information in the User-Agent header field, as well as the navigator.userAgent, navigator.appVersion, and navigator.platform JS APIs.

Once this is complete, you will still be able to reliably get the browser major version, platform name, and distinguish between desktop and mobile (or tablet), solely from the User-Agent string. For more advanced use cases, you should migrate to the User Agent Client Hints API.

Note: We have no plans to change the User-Agent string on Android WebView or Chrome for iOS at this time, but will make public updates if and when that changes.

Our current high-level plan is as follows:

  • Beginning in M92, we plan to start sending deprecation notices for the navigator.userAgent, navigator.appVersion, and navigator.platform getters in the DevTools Issues tab.
  • In the coming weeks, we will announce an Origin Trial for sites to opt in to receiving the fully reduced User-Agent. We expect to run the Origin Trial for at least 6 months to provide enough time for sites to opt in, test, and provide feedback on the feasibility and compatibility of our desired end state.
  • We will evaluate feedback from Origin Trial partners and the community, and based on this feedback proceed to Phases 3 through 7 of our plan (see next section for details), giving the ecosystem adequate time to adapt in between them. Otherwise, depending on feedback we will reconsider the best course of action.
  • For sites with complex use cases that require more time for migration, we aim to offer the ability to extend the current User-Agent behavior for at least an additional 6 months (through a "reverse Origin Trial").
Proposed rollout plan

We plan to roll out these changes slowly and incrementally in 7 Phases—pending Origin Trial feedback—and plan to publish an update soon on the proposed timing and milestones beyond Phase 1.

Reduction Preparation

Phase 1: Warn about accessing navigator.userAgent, navigator.appVersion, and navigator.platform in DevTools, beginning in M92.

Phase 2: Launch an Origin Trial for sites to opt into the final reduced UA string for testing and feedback, for at least 6 months.

Reduction Rollout

Phase 3: Launch a reverse Origin Trial, for instances where a site may need more time for migration, for at least 6 months.

Phase 4: Ship reduced Chrome MINOR.BUILD.PATCH version numbers (“0.0.0”). Once rolled-out, the reduced UA string would apply to all page loads on desktop and mobile OSes that do not opt into the reverse Origin Trial.

Phase 5: Begin roll-out of reduced Desktop UA string and related JS APIs (navigator.userAgent, navigator.appVersion, navigator.platform). Once rolled-out, the reduced UA string would apply to all page loads on desktop OSes that do not opt into the reverse Origin Trial.

Phase 6: Begin roll-out of reduced Android Mobile (and Tablet) UA string and related JS APIs. Once rolled-out, the reduced UA string would apply to all page loads on Android that do not opt into the reverse Origin Trial.

Reduction Completion

Phase 7: reverse Origin Trial ends and all page loads receive the reduced UA string and related JS APIs.

See the companion Reduced User Agent string updates page for more details and example User Agent strings at each of these phases.

What do I need to do to get ready as a developer?

Our plan was designed with backwards compatibility in mind, and while any changes to the User Agent string need to be managed carefully, we expect minimal friction for developers as we roll this out (i.e., existing parsers should continue to operate as expected).

If your site, service, library or application relies on certain bits of information being present in the User Agent string such as Chrome minor version, OS version number, or Android device model, you will need to begin the migration to use the User Agent Client Hints API instead.

If you don’t require any of these, then no changes are required and things should continue to operate as they have to date.

Why are we doing this?

As noted in the User Agent Client Hints explainer, the User Agent string presents challenges for two reasons. Firstly, it passively exposes quite a lot of information about the browser for every HTTP request that may be used for fingerprinting. Secondly, it has grown in length and complexity over the years and encourages error-prone string parsing. We believe the User Agent Client Hints API solves both of these problems in a more developer- and user-friendly manner.

What about other browsers?

In some ways Chrome is playing catch up on this front: Safari was the first to cap the macOS version number in the UA string and Firefox has followed suit. Firefox has also capped the Windows version number to 10.

Learn More

Posted by Mike Taylor and Jade Kessler, Chrome Team

Back in February, we announced that cross-origin isolation will be required on all platforms in order to access APIs like SharedArrayBuffer and performance.measureUserAgentSpecificMemory() starting in Chrome 91. Based on your feedback and issues reported, we've decided to adjust the timeline for SharedArrayBuffer usage in none cross-origin isolated sites to be restricted in Chrome 92.

Your feedback is important and we are listening.

Posted by Lutz Vahl, Technical Program Manager