4.x seems to be significantly slower (not related to brotli!)

[indeyets]

I see that my app works with much higher CPU load and produces 20% throughput at the same time.

3.1.0 with no options: fast (119MB/s)
4.0.0 / 4.0.1 with no options: slow (20 MB/s)
4.0.x with { br: false }: slow
4.0.x with { br: false, gzip: false, deflate: false }: fast again

same machine, same nodejs version (12.16.3)

3.1.0 produces compressed output (I checked)

trying to trace it further

[indeyets]

@uhop can it be related to the new "external" checks of Accept: headers?

[uhop]

4.0.0 was based on my ideas, but not on my code. I have to look first and compare versions.

In general, I would assume that any checks are fast, and, probably, done just once, and all CPU time is spent compressing. Could it be that gzip is running with a higher compression rate by default than in 3.1.0? One way to see it is to compare sizes produced by different versions.

I'll investigate.

[indeyets]

Could it be that gzip is running with a higher compression rate by default than in 3.1.0

I doubt it is, as default invocation is without any options both in old and in new code. Will try to run another round of tests a bit later

[indeyets]

ok. I guess it is related to #112

benchmark tool does not request compressed responses by default, but it gets one now.

I'll make separate investigation about why it hurts my production installation so hard.

I will close this issue, but #112 should be fixed