search

davidmarkclements / keepings-node.js-fast

Repository accompanying the Keeping Node.js Fast article
31 stars 9 forks source link

How clinic profiling is reliable? #1

Open ghost opened 6 years ago

ghost commented 6 years ago

First, thanks for writing such a good post! :+1:
I've been trying to reproduce the same results you described in the post. Even though I know the "fast" version is way better (I can see the stats from autocannon, they are indeed faster) when I generate any report from clinic I can't conclude anything. If I run just with the -c100 option I don't see any event loop issue, so I pushed harder and tried with -c 1000 -p 10 and I have event emitter and memory issues on both versions. Same with flame graphs. I don't see any block stack redder than "cold" color. So I'm wondering if I did something wrong or if the profiling is just not reliable. (Let me know if you want to see the pics, I can attach here)

meganmcmillan commented 5 years ago

I would like to second the thanks for such a good post! I am running into the same issue as @chicocode when trying to reproduce your results. If you have any insight on this issue, it would be much appreciated!