just VS picohttpparser

just

https://github.com/just-js/just Is another for a V8 runtime, it really shows how well optimized it is. https://www.techempower.com/benchmarks/#section=data-r21

i just tried recompiling v0.0.2 (https://github.com/just-js/just/releases/tag/0.0.2) of just-js and comparing it to current. for the completely static build on ubuntu 22.04 i see following:

0.0.2 (v8 v8.4.371.18) - file size: 15.2 MB, startup RSS: 8.4 MB

Not at all. Javascript, as well as java compiled into the bytecode, but just incrementally and at the runtime. You cannot compile typescript into bytecode directly (at least it intend to be like that). You can even compile js to executable (https://github.com/just-js/just). So no, typescript transpiles to javascript, whereas scala compiles to bytecode, it's different things

Trust me, these guys are insane and go to really extreme levels and use optimization techniques which are not generally prevalent among the general programming fraternity. for eg, look at pico.v it is awesome and just-js is truly unbelievable, then there is faf, most of them combine low level programming trickery to reach those insane numbers. Also some of the code may not be useful in a production app but they actually extract the juice out of the metal at every instance. Memory optimizations, compiler optimizations, postgresql wire implmentations, rust black magic, these guys are really crazy and passionate.

Yea, it is definitely a fake HTTP server which I acknowledge in the article [1]. However based on the size of the requests, and my observation of the number of packets per second being symmetrical at the network interface level, I didn't have a concern about doubled responses.

Skipping the parsing of the HTTP requests definitely gives a performance boost, but for this comparison both sides got the same boost, so I didn't mind being less strict. Seastar's HTTP parser was being finicky, so I chose the easy route and just removed it from the equation.

For reference though, in my previous post[2] libreactor was able to hit 1.2M req/s while fully parsing the HTTP requests using picohttpparser[3]. But that is still a very simple and highly optimized implementation. From what I recall when I played with disabling HTTP parsing in libreactor I got a performance boost of about 5%.

1. https://talawah.io/blog/linux-kernel-vs-dpdk-http-performanc...

2. https://talawah.io/blog/extreme-http-performance-tuning-one-...

3. https://github.com/h2o/picohttpparser

Just-js is not nodejs framework. It's sperate runtime and most of the http code is written using c/c++ (for example headers parsing logic is written using c and is using https://github.com/h2o/picohttpparser which is c library)