Chrome 120

Chrome 120 results are published.

Attached you find a binary build of all frameworks such that you can run the benchmark without installing and building all frameworks (see README for how to use the build.zip)

There's a new view for compressed and uncompressed transfer size and the first paint metric. I'm planning to remove lighthouse metrics so please post in #1492 if you want me to keep it.

The standard view shows the total duration for each benchmark (i.e. including javascript/wasm and rendering duration). There's a new JS only view that shows script duration only and a paint duration only mode. Both are still experimental please post feedback to #1233

The append to table benchmarks changed a bit in this benchmark: It includes now warmup runs.

Chrome 119

Chrome 119 results are published.

Overall results appear to be comparable with chrome 118. Notable exceptions are:

• blazor got faster with chrome 119 (but still one of the slowest)
• react redux got slower (we saw that in #1469) already.

Attached you find a binary build of all frameworks such that you can run the benchmark without installing and building all frameworks (see README for how to use the build.zip)

Some frameworks were archived after the chrome 118 run, in case you're missing some implementation.

Since chrome 118 we're now using a weighted geometric mean for the overall result
The number of runs for the CPU benchmarks was bumped from 10 to 15 for all benchmarks except select row, which was bumped to 25 runs. We hope this improves accuracy of the benchmarks

Chrome 118

Chrome 118 results are published.

Please note that there have been significant changes for this run:

• We're now using a weighted geometric mean for the overall result
• The number of runs for the CPU benchmarks was bumped from 10 to 15 for all benchmarks except select row, which was bumped to 25 runs. We hope this improves accuracy of the benchmarks

So please don't hesitate to post in the discussion sections if you find something suspicious.

Attached you find a binary build of all frameworks such that you can run the benchmark without installing and building all frameworks (see README for how to use the build.zip)

Chrome 117

Chrome 117 results are published. Results look pretty close to 116.

Attached you find a binary build of all frameworks such that you can run the benchmark without installing and building all frameworks (see README for how to use the build.zip)

Chrome 116

First of all: I don't like chrome's early stable updates and the fact that I can't opt in. So I decided to run official runs against the chrome beta version. I'm trying to pick the same version number as the early stable update, which is 116.0.5845.82 for chrome 116.

Nevertheless the official results have been updated: Chrome 116 results are published. I haven't found significant changes from chrome 115.

Attached you find a binary build of all frameworks such that you can run the benchmark without installing and building all frameworks (see README for how to use the build.zip)

Chrome 115

Chrome 115 results are published. Looks like partial update got a little faster.

Attached you find a binary build of all frameworks such that you can run the benchmark without installing and building all frameworks (see README for how to use the build.zip)

Chrome 114

Chrome 114 results are published. A nice release! it looks significantly faster than chrome 113 (only except for select rows)

Attached you find a binary build of all frameworks such that you can run the benchmark without installing and building all frameworks (see README for how to use the build.zip)

Chrome 113

Chrome 113 is out - welcome WebGPU!

As far as the benchmark is concerned some benchmarks like create rows got a tiny bit slower. On the other hand partial update got significantly faster (but it has a 16x slowdown, so it could be either that chrome is faster or that the cpu is less throttled...)

There were some major changes for the test driver. I'm trying to measure both the total duration and the JS script execution duration, but results are not ready to be published yet. This means that the structure of the files in the results directory changed and is not compatible with older versions (this also applies to the results.json and results.ts files).

Attached you find a binary build of all frameworks such that you can run the benchmark without installing and building all frameworks (see README for how to use the build.zip)

Chrome 112

Results for Chrome 112 are ready: https://krausest.github.io/js-framework-benchmark/2023/table_chrome_112.0.5615.49.html

Looks very similar to Chrome 111. Both vanillajs and vanillajs-1 deserve getting updates such that they are fastest again in all benchmarks 😄 . PRs would be welcome...

Attached you find a binary build of all frameworks such that you can run the benchmark without installing and building all frameworks (see README for how to use the build.zip)

Chrome 111

Results for Chrome 111 are published: https://krausest.github.io/js-framework-benchmark/2023/table_chrome_111.0.5563.64.html

Looks very similar to Chrome 110.

Attached you find a binary build of all frameworks such that you can run the benchmark without installing and building all frameworks (see README for how to use the build.zip)