JavaScript vs Wasm: Which is More Energy Efficient and Faster?

Which is faster and more energy efficient, JavaScript or Web Assembly? The University of Minho in Portugal studied exactly this question and found that JavaScript is more energy efficient and faster than Wasm when it comes to in-lab microbenchmarks. , but in real-world applications, we’ve come to the conclusion that Wasm beats JavaScript in terms of speed and energy efficiency — by as much as 30% on average in some cases.

And it’s still in early Wasm.

“Wasm is still in its early stages and time will tell how it evolves,” study author and software engineer João De Macedo told New Stack in an email. “Wasm pushing native apps out of the operating system entirely and pushing the web browser into his 21st century operating system seat is our boldest prediction.”

Microbenchmark vs. real application

The study, published in 2022, looked at real-world and micro-benchmarks.

“A microbenchmark is a program that tracks and measures the performance of a single, well-defined task, such as elapsed time, operation rate, or bandwidth,” explains De Macedo. “Microbenchmarks are one of the leading ways to measure the performance of software systems, and Wasm is no exception.

“Comparing the execution time and energy performance of Wasm and JS is particularly important,” he added, because one of Wasm’s main goals is to improve the performance of web applications.

Enlarged image.

Regarding microbenchmarks, JavaScript can be more energy efficient and faster than Wasm in some cases. Wasm was energy efficient and outperformed JavaScript on Google Chrome and Microsoft Edge, but JavaScript showed better performance results than Wasm on Mozilla Firefox, “most of the time with a big difference.”

Ultimately, however, Wasm still ruled the world in real applications.

“Preliminary results show that WebAssembly is still in its early stages, but has already begun to surpass JavaScript and has room for further growth,” said the report. “Statistical analysis shows that WebAssembly has significant performance differences compared to JavaScript.”

How the JavaScript vs. Wasm Survey Works

In this study, we used the Wasmboy benchmark, a Gameboy/Gameboy Color emulator written in Typescript to benchmark Wasm, to see how well Wasm and JavaScript perform in real-world applications. Wasmboy is written in JavaScript/TypeScript and was created with the primary purpose of comparing runtime performance between Wasm generated by the AssemblyScript compiler and his ES6 latest version of JavaScript generated by the TypeScript compiler. rice field.

“This game console includes six different open source games that can be run on the console,” the report said. “We’ve updated WasmBoy’s source code to specify which browser the game should run in.”

So we ran 6 games in 2 languages ​​in 3 browsers: Chrome, Edge, and Firefox, so the team had a total of 36 unique samples.

I also used the PSPDFKit benchmark. It uses software that supports form viewing, annotating, and filling PDF documents on any platform. The report notes that an open source benchmark was created to assess the feasibility of porting software to the Wasm ecosystem and comparing Wasm and JavaScript implementations. The team made changes to the application source code and performed some typing in his two languages ​​considered (was and asm.js). To run the benchmark with realistic inputs, the team considered five different PDF documents of his. This includes one of his books divided into three parts, one scientific paper, and a slide presentation consisting of 20 slides.

“Similar to the Wasmboy benchmark, we have developed a makefile to automate execution in different browsers,” he says, obtaining five samples in each language to run 30 unique programs in three browsers. the report said.

I also looked at various microbenchmarks. These programs were originally written in C and compiled to both Wasm and JavaScript using the Emscripten compiler. Other languages ​​that can be used to compile to Wasm are C/C++, Rust, Go, Python, and AssemblyScript in the form of TypeScript.

You can read more about the microbenchmarks and other aspects of the research in “WebAssembly vs JavaScript: Energy and Runtime Performance” by João De Macedo, Rui Abreu, Rui Pereira, and João Saraiva. Free with Institute of Electrical and Electronics Engineers.org or De Macedo master’s thesis.

But in short, this approach allowed the team to explore how scale and input size, for example, are handled differently in JS and Wasm. Previous research used this approach, but only looked at virtual machine performance. De Macedo’s research required examining real-world applications, so the team developed a framework for measuring performance within a browser-based environment.

Did you use Wasm as an alternative to JavaScript?

I asked De Macado if Wasm will eventually replace JavaScript.

“Probably not, because JS is better suited for web pages that don’t need to be ultra-high performance,” he replied. “Currently, Wasm serves as a complement to his JS rather than a replacement. However, if Wasm is further developed, it will have faster load times and more efficient resource usage, so some applications he may replace JS.”

In the long run, De Macado predicted Wasm would be disruptive.

“Wasm isn’t just going to revolutionize the web,” he said. “The potential of WebAssembly has the potential to upend multiple areas of the technology market, including the cloud.In recent years, there has been an increasing shift to a container model that may not always meet the needs of all organizations. I have.”

Loraine Lawson is a veteran technology reporter who has covered technology issues from data integration to security for 25 years. Prior to joining The New Stack, she was the editor of banking and technology site Bank Automation News. she has her…

Read more about Lorraine Lawson