Introduction
Web browser performance evolves rapidly, and benchmarks must keep pace. JetStream 3.0, developed jointly by WebKit, Google, and Mozilla, represents a major leap forward—especially in how WebAssembly (Wasm) workloads are measured. This guide walks you through the key innovations behind JetStream 3, explaining why older benchmarks fell short and how the new suite solves critical problems like the infamous “infinity score.” By the end, you'll understand the engineering philosophy driving modern web performance testing.
What You Need
- A basic understanding of web performance concepts (page load time, JavaScript execution)
- Familiarity with WebAssembly (optional but helpful)
- A modern web browser (Chrome, Firefox, or Safari) to run JetStream 3
- Access to the JetStream 3 benchmark site (official link)
- About 30 minutes to read through this guide and explore the suite
Step 1: Recognize the Limitations of Legacy Benchmarks
Benchmarks are essential tools for browser engine developers, but they have a shelf life. The web constantly adopts new best practices, and a benchmark that once accurately reflected real-world performance can quickly become outdated. As the most obvious optimizations are exhausted, engineers begin fine-tuning specifically for the benchmark’s artificial workloads—making improvements that don’t translate to actual web applications. JetStream 3 was born from the need to refresh the measurement framework and shift focus toward more realistic, scalable workloads.
Step 2: Understand the WebAssembly Startup Puzzle
When JetStream 2 was released, WebAssembly was still emerging. The expectation was that Wasm would be used for large C/C++ applications (like games) where users accepted long startup times in exchange for high throughput. Consequently, JetStream 2 split Wasm scoring into two phases: Startup and Runtime. But as browser engines optimized instantiation times—WebKit reduced startup to nearly zero for smaller modules—the benchmark’s timer (Date.now()) began rounding sub‑millisecond times to 0 ms. The scoring formula Score = 5000 / Time then produced infinity. This “infinity problem” forced a patch in JetStream 2.2 to cap scores at 5000, but it highlighted a deeper issue: the benchmark's structure no longer reflected real Wasm usage patterns.
Step 3: Learn How JetStream 3 Redesigns Wasm Measurement
JetStream 3 doesn’t just patch the old approach—it transforms the measurement philosophy. Instead of treating startup and runtime as separate, artificial phases, the new suite integrates WebAssembly into more complex, realistic workflows. Wasm is now used for image decoders, UI frameworks, and libraries that are part of the critical page load path. By eliminating the separate startup timer, JetStream 3 avoids the infinity problem entirely and ensures that optimizations benefit actual user experiences. The shift also encourages engine developers to focus on end‑to‑end performance rather than micro‑optimizing a single phase.
Step 4: Explore the Collaborative Engineering Behind the Suite
JetStream 3 is the result of extensive cooperation between the WebKit, Google, and Mozilla teams. Each group contributed benchmarks reflecting real workloads from their browsers. The WebKit team, for example, invested heavily in JavaScriptCore optimizations—such as faster Wasm instantiation and smarter compilation—that directly influenced the suite’s design. Understanding this collaborative process helps you appreciate why JetStream 3 is more robust and representative than any single‑vendor benchmark.
Step 5: Apply JetStream 3 Insights to Your Performance Analysis
Now that you know the benchmarks’ rationale, use JetStream 3 to evaluate browser performance. Run the full suite, and pay particular attention to Wasm‑related subtests. Compare scores across browsers and versions. Because the tests now mirror real‑world usage, improvements in your JetStream 3 score will likely translate to faster page loads, smoother interactions, and better responsiveness in applications that rely on WebAssembly.
Tips for Using JetStream 3 Effectively
- Run multiple trials: Close other applications and run the suite at least three times to account for variance. Use the median or average score.
- Focus on subtests, not just the aggregate: A high overall score might hide weaknesses in specific areas. Examine each test’s performance to pinpoint where a browser excels or lags.
- Keep your browser up to date: JetStream 3 targets modern engines. Outdated browsers may not support all features or may penalize performance unfairly.
- Understand the scoring scale: JetStream 3 uses a geometric mean of subtest scores normalized to a reference browser. A score of 100 means the system is as fast as the reference; higher is better.
- Use the results to drive real improvements: If a particular subtest (e.g., Wasm image decoding) underperforms, investigate why. It might point to a bug or missing optimization in the engine.