HN Explorer

3 comments

[ 3.3 ms ] story [ 15.0 ms ] thread
jlongr
In the first example, the JIT version takes 0.218s and the non-JIT version takes 0.0007s.

Isn't the JIT version supposed to be faster?

jjuel
Yeah that is interesting. I hadn't noticed on my first read through. He was absolutely touting the benefits of JIT with a claim "In some cases a JIT version could make code run thousands of times faster, but benchmarking is key." I guess this benchmark shows that it is quite a bit slower in this particular case.
falsedan
Isn’t all the performance improvements by JIT completely swamped by the abysmal startup cost of the python executable?