Stefan Krause

@robbiespeed good catch, indeed ivi stands out that the dom nodes created by initBenchmark are still reported by chrome (in contrast to vanillajs, doohtml and solid). @localvoid is that intentional?

@robbiespeed Maybe we're really on to something Seems like forceGC() which called window.gc() a few times didn't clean the dom nodes for ivi. If we run window.gc() and HeapProfiler.collectGarbage" the...

Here's the result with the modified forceGC method: Looks much more reasonable to me: * Adding a delay doesn't yield a different order * Much closer to manual testing

With 1.000 msecs delay the graph looks pretty similar, which is really good news!

Conclusion: * Somehow calling 7x window.gc() didn't clean up the dom nodes for ivi. * Running window.gc() and HeapProfiler.collectGarbage in a loop with seven iterations clears the dom nodes for...

I'm sorry, all PRs are currently blocked by https://github.com/krausest/js-framework-benchmark/issues/1655

Thanks - I hope to publish results for the PR tomorrow.

[Results](https://krausest.github.io/js-framework-benchmark/current.html) are updated. Nice first run!

I'm sorry, all PRs are currently blocked by https://github.com/krausest/js-framework-benchmark/issues/1655. Feel free to submit a PR for #1661 to get this PR merged 😄

Thanks - I hope to publish results for the PR tomorrow.