🏠IT之家•Stalecollected in 12m
Chrome Browser Hits New JetStream 3 Benchmark Record
💡Performance gains in Chrome's V8 engine directly impact the speed of web-based AI tools and LLM frontends.
⚡ 30-Second TL;DR
What Changed
10% performance boost in JetStream 3 benchmark
Why It Matters
Faster browser performance directly benefits web-based AI applications and LLM interfaces, reducing latency for complex client-side computations.
What To Do Next
Benchmark your web-based AI models on the latest Chrome version to leverage improved WebAssembly performance.
Who should care:Developers & AI Engineers
Key Points
- •10% performance boost in JetStream 3 benchmark
- •Optimized V8 compiler for faster asynchronous operations and string comparisons
- •Enhanced WebAssembly support for AI and cryptographic workloads
- •Blink engine improvements in style calculation and DOM manipulation
🧠 Deep Insight
Web-grounded analysis with 21 cited sources.
🔑 Enhanced Key Takeaways
- •JetStream 3 is a collaboratively developed benchmark by major browser vendors including Apple, Google, and Mozilla, designed to measure the full lifecycle performance of advanced web applications, encompassing startup, worst-case, and average-case execution rather than isolated phases.
- •The benchmark significantly expanded its WebAssembly (Wasm) coverage, now constituting 15-20% of the suite (up from 7% in JetStream 2), to better reflect the growing importance of Wasm in modern web applications, particularly for AI and cryptographic workloads.
- •Beyond the JetStream 3 record, Chrome also achieved a record score of 61 on Speedometer 3.1, marking a 5% improvement since its last benchmark update approximately a year prior, indicating broader performance enhancements across various browser metrics.
- •V8 engine optimizations included inlining 'fast paths' for common JavaScript operations, with a significant impact on asynchronous tasks like microtask dispatch and await resolution, as well as improvements in sorting data and string comparisons.
- •Blink rendering engine improvements involved implementing smarter caching mechanisms, such as a unified
querySelector()cache, streamlining lookup paths for style resolution and DOM interactions, and leveraging SIMD operations for more efficient string copying and HTML parsing.
🛠️ Technical Deep Dive
- V8 JavaScript Engine Optimizations:
- Asynchronous Operations: Implemented inlining of "fast paths" for microtask dispatch and await resolution, allowing the optimizing compiler to skip unnecessary execution steps for common asynchronous cases.
- String Handling: Applied similar fast-path inlining techniques to sorting data and comparing strings, yielding significant speed gains.
- BigInt Handling: Improved BigInt division and canonicalization, along with changes to data structure allocation to enhance security within the sandbox.
- WebAssembly Integration: Reduced overhead for JavaScript-to-WebAssembly function calls by eliminating repeated type conversions and loads within loops.
- Compiler Architecture: V8 utilizes a tiered execution engine comprising Ignition (an interpreter), Maglev (a fast optimizing tier), and TurboFan (an aggressive optimizing compiler), which continuously analyzes and optimizes code.
- Object Optimization: Relies on "hidden classes" and stable object shapes to enable efficient property lookups and inline caching, improving performance by treating dynamic JavaScript objects as if they had fixed structures.
- WebAssembly (Wasm) Enhancements:
- Core Optimizations: Optimized V8's internal data structures, SIMD (Single Instruction, Multiple Data) instructions, and register allocation specifically for AI, cryptography, and interpreter use cases.
- Memory Management: Enhanced memory reuse during compilation processes.
- WasmGC Performance: Improved WasmGC (Garbage Collection) performance through load and null-check elimination and speculative inlining of indirect function calls with deoptimization support.
- Workload Coverage: JetStream 3's Wasm workloads now include benchmarks for Rust, Kotlin/Wasm, Dart2wasm, J2CL, and .NET toolchains, and test features like WasmGC, SIMD, and Exception Handling.
- Wasm 3.0 Features: The WebAssembly 3.0 specification, released in 2025, introduced garbage collection, 64-bit memory addressing (Memory64), and exception handling, allowing managed languages to target Wasm more efficiently.
- Blink Rendering Engine Improvements:
- DOM API Response: Implemented smarter caching and streamlined lookup paths for style resolution and DOM interactions.
- Unified Caching: Deployed a unified
querySelector()cache across the rendering engine to minimize redundant DOM lookups. - Attribute Handling: Optimized element attribute handling by introducing a fast bailout path that skips unnecessary checks when attributes are absent.
- Resource Caching: Began caching dataURI resources beyond the scope of a single document, benefiting modern web frameworks.
- Style Recalculation: Reduced style recalculation delays for transitions and simplified CSS selector caching and matching to maintain fluid animations.
- Text Processing & HTML Parsing: Leveraged SIMD operations to make string copying significantly more efficient and optimized string creation directly within the HTML fast-path parser to reduce overhead in constructing DOM elements.
- Typography & Graphics: Resolved performance bottlenecks in typography and vector graphics rendering, including Apple Advanced Typography (AAT) shaping.
- Style Invalidation: Uses "Invalidation sets" to store metadata of selectors in stylesheets, acting as a mutation filter to identify and mark only potentially affected elements for restyle, rather than recalculating everything.
🔮 Future ImplicationsAI analysis grounded in cited sources
The performance gains will accelerate the development and adoption of highly complex, compute-intensive web applications, particularly in AI and data processing.
Optimizations specifically targeting AI and cryptographic workloads in WebAssembly, coupled with overall JavaScript execution improvements, directly support the execution of advanced applications directly within the browser.
The collaborative nature of JetStream 3 and Chrome's record will intensify competition among browser vendors, leading to further web platform performance enhancements.
Benchmarks like JetStream 3 are explicitly designed to foster 'healthy competitiveness' among browser engines, driving continuous optimization across the web ecosystem.
Developers will increasingly leverage WebAssembly for critical performance sections of web applications, reducing reliance on JavaScript for heavy computational tasks.
The significant increase in WebAssembly coverage in JetStream 3 and ongoing Wasm optimizations, including WasmGC and Memory64, make it a more robust and performant target for high-performance code.
⏳ Timeline
2008-09
Google Chrome browser launched, with speed as a foundational principle.
2018-09
Chrome's V8 engine performance improved 20x over 10 years, introducing the Ignition interpreter and TurboFan compiler, and WebAssembly with the Liftoff compiler.
2022-03
Chrome achieved over 300 on Apple's Speedometer 2.0 benchmark on the M1 MacBook, marking the highest score seen at that time.
2024-06
Chrome achieved the highest score ever on Speedometer 3.0, demonstrating a 72% increase in its Speedometer score since May 2022.
2026-01
WebAssembly 3.0 was announced, standardizing 11 new features including Garbage Collection and 64-bit memory addressing (Memory64).
2026-03
JetStream 3.0 benchmark suite was officially released, developed collaboratively by major browser engines including Apple, Google, and Mozilla.
2026-06
Google Chrome achieved a new record score of 469 on JetStream 3 and 61 on Speedometer 3.1, measured on a MacBook Pro M5 running MacOS 26.0.1.
📎 Sources (21)
Factual claims are grounded in the sources below. Forward-looking analysis is AI-generated interpretation.
📰
Weekly AI Recap
Read this week's curated digest of top AI events →
👉Related Updates
AI-curated news aggregator. All content rights belong to original publishers.
Original source: IT之家 ↗