Spectre & Meltdown Patches vs WWW

Some 20 days ago Greg Rozmarynowycz here at SO asked

What are the performance impacts on web services resulting from Meltdown patches?

He probably knew, as I know, that this kind of question is not ideal for SO for many reasons. Since he did asked, I am going to ask as well.

He made me search for how M&S patches effect JS engines, common bundlers such as Webpack and tech such as WebGL. You see, I am more interested in client side than in cloud. Since there are JS exploits related to M&S and some web apps are saucing CPU/GPU extensively, the question seems to be appropriate. But I have not found any conclusive and comprehensive set of benchmarks or studies targeting this field. I have tried to measure time vs CPU power during build of Angular5 by Angular CLI 1.6 before and after updating my machines and also the performance of final app in production in evergreen browsers. There seems to be a difference. But I would not put my hand in fire for what I did. I am no expert to say and I have no appropriate tools to run any conclusive tests (I do not even know how to do it scholarly).

Did anybody tried to investigate these issues? Does it make a sense to investigate them? What is the appropriate way of benchmarking JS processes in browser? Is jsperf.com a reliable source of JS-related test cases? Are there any solid data on this or is just everybody keep saying how bad these patches are and how largely they slow down everything?