Yeah it's built on the quickJS runtime that has a low footprint and fast startup but no JIT, so in anything longer than the v8 startup, it ends up being about a magnitude slower

[–] RonSijm@programming.dev 1 points 9 months ago (1 children)

https://github.com/awslabs/llrt/raw/main/benchmarks/llrt-ddb-put.png
https://github.com/awslabs/llrt/raw/main/benchmarks/node20-ddb-put.png

Maybe I'm just stupid, but what are these numbers?

"HTTP benchmarks measured in round trip time for a cold start"

Soo, I'm guessing it's round trip time in milliseconds?

What is p0 to p100? Are they putting 0 to a 100 items? Are they putting 1 item into a dataset of size p..?

[–] keccsx@programming.dev 1 points 9 months ago* (last edited 9 months ago)

Percentiles of the measured startup times.

p100: 100% of the startups were slower, this is the maximum time they measured.

p50: 50% were slower than this

And so on to p0, which is the smallest measured time.