Are We Fast Yet

Recently, two colleagues and I spent some time creating a website: Are We Fast Yet.

The name and layout of the site are inspired by this site, which was set up 14 years ago by members of the Firefox team to track performance gaps between their browser and others¹.

Additionally, considering the Mozilla team’s tradition of sending a cake to the Internet Explorer team every time a new version was released², I see this as a great cultural example of how teams can remain friendly and respectful even while competing.

After spending two years in the blockchain industry, I’ve found the relationships between developer teams here to be quite strange. Sometimes they quietly and quickly “borrow” code from each other, keeping it secret to prevent more people from discovering the valuable resource. Other times, they proudly claim to be the best without any reliable data to support their claims.

I still can’t understand why, in this industry, the harder I seek the truth, the more chaos I encounter. If it weren’t for my basic common sense reminding me that web3 developers and I live on the same planet, I’d be tempted to believe I slip into a different metaverse for a few hours each day.

In comparison, it’s easy to see why I favor Ethereum, especially the core developer team. If the web3 software industry is a vast desert, Ethereum is like a fragile oasis. If it’s a collection of potentially habitable exoplanets, Ethereum is like Mars right now.

The reason for creating this website is that I feel there’s something missing in the ethos shared by the Ethereum core team: a fear of open competition. As a client user and software developer, I can clearly see the significant performance differences between clients. Isn’t this one of the top priorities for a software developer to address?

Despite the ongoing rhetoric about “all clients matter,” we continue to play emotional and politically correct cards. Why can’t we directly play the performance card? Just like Chrome once did with its “Faster than Potato/Sound/Lightning” campaign to attract more users.

I understand why it’s missing. The relationships among these teams are a bit delicate; they aren’t direct competitors since they’re all somewhat funded by the same ecosystem. In contrast, Firefox vs. Chrome had more independent identities.

But no worries, Endurance doesn’t have this problem. We have no direct financial ties with the Ethereum Foundation. We’re not funded by them, nor are they by us. We have a purer, more impartial stance to ensure these clients reach their performance potential. Interestingly, we’re not just a “third-party testing agency” because our network consists of these clients, and their performance affects our network’s performance. So I believe we’re well qualified to say, “Are We Fast Yet.”

Lastly, I want to briefly introduce the logic behind our initial test suite selection.

On our initial website launch, we track three test suites, two of which come from Ethereum core client teams.

• BurntPix Benchmark, provided by Geth developer Péter Szilágyi. According to the current test results on the website, Geth isn’t the best.
• Gas Benchmarks, provided by the Nethermind team. Similarly, Nethermind isn’t the fastest in this suite either.

So, I believe these test suites are reliable and not biased toward their teams’ clients.

Without these two test suites, the website couldn’t have launched. Many thanks to them.

The last test suite is from our team. Since our network went through an upgrade and carries historical state baggage, our Genesis file is particularly large, so we also focus on this area of performance.

In the future, our team will gradually add more comprehensive test suites with our limited bandwidth. For example, the parallel processing speed of hundreds of transactions in one block? Or the synchronization speed of a miniature Archive node?

Also, we welcome everyone to provide us with more test suite leads (both EL/CL) and to submit ISSUES/PRs to our repo to optimize the testing framework.

For more questions, please visit the repo here for discussion.