alleonhardt
commented
3 years ago I hope you have read the benchmark code techempower used for this benchmark. The template they tested is this one:
<!DOCTYPE html>
<html>
<head><title>Fortunes</title></head>
<body>
<table>
<tr><th>id</th><th>message</th></tr>
<% for item in items { %><tr><td><%= item.id %></td><td><%= &*item.message %></td></tr><% } %>
</table>
</body>There are also only 12 entries inside the database table they use to instantiate the template. The second RUST ONLY benchmark you bring up inserts 100 entries. I dont know which one is faster and I also dont want to judge, but math is not comparing two completely different benchmarks with each other. If you care that much about speed just benchmark the libraries in your use case and use the one that's fastest for your particular case.
The second point is the benchmark you quote, tests all kind of framworks written in Rust but also written in C, C++ etc. There are only two rust frameworks that perform better than sailfish and both only by a small margin.
I think it is very rude to insult someone who created a great library which you can use for free in his/her free time.
d4h0
zzau13
Kogia-sima
https://www.techempower.com/benchmarks/#section=test&runid=646af5ea-0fae-448d-b77d-bc27b13d73e5 While in your benchmarks and these https://github.com/djc/template-benchmarks-rs , they said otherwise. It would be very honorable to write an article explaining why to the community, which I think has the right to know why you lie in this way as reflected in the results of techempower.
It really is outrageous and an insult to me that you had put those results.
And remember math doesn't fail, don't bet against it.