Elm 0.17 isn't that fast with TodoMVC Benchmark v1.0.0

[D_K]

Can anyone comment about Elm code in the benchmark?
https://developit.github.io/preact-perf

[D_K]

Results with less tests at once:

[gampleman]

1. That’s not really that bad. Being about 20% faster than the industry standard framework isn’t a bad result. How much more you can optimise out of a simple synthetic benchmark isn’t really that important in practice.

2. I’d take a relatively trivial benchmark like this with a big pinch of salt. First of all the benchmark ranking seems pretty variable (even your own two screenshots show wildly different rankings; mine are different still). That’s not a good sign for a solid benchmark suite aiming to compare “frameworks”.

3. Elm 0.17? Who cares? Do you still use Elm 17? Elm 19 came out 4 years ago…

[D_K]

1. I wouldn’t call operations on list a synthetic benchmark. Not everyone uses them, I see point in this benchmark.

2. I agree that results vary right now and I see even 200-300% performance spikes between runs. They should work hard with consistency of benchmarks.

3. I have no opinion. This is what Google brings with “TodoMVC Benchmark” search.

[pdamoc]

Looks reasonable on my machine:

Later Edit: the results of the benchmark are not really reliable. This becomes obvious if one runs the suite multiple times.