Fork me on GitHub
#luminus
<
2016-01-26
>
yogthos03:01:05

@gerred @shanekilkelly I'd be definitely interested in figuring out if there's a way to squeeze more performance out

yogthos03:01:05

I've just recently updated to the latest version for the next round, and there have been a lot of changes since the last one

yogthos03:01:45

I've tuned up logging, and switched to using hikaricp, so that should definitely help

yogthos03:01:18

otherwise there really shouldn't be that much difference with using plain compojure, the default middleware might also be a factor

yogthos03:01:41

for example, everything goes through wrap-formats middleware and that could be turned off

yogthos03:01:23

I'm not sure that would be representative of real world use though, so depends whether the goal is to get the best performance possible or to illustrate likely baseline performance

gerred05:01:26

@yogthos makes sense. It was fairly surprising. By the way, in this note, happy to help review your second edition, I have a copy of the first.

yogthos05:01:39

great, the beta release should be on the 2nd

yogthos05:01:04

and things are looking to be in good shape, since I missed the last window I had a bit more time to polish things simple_smile

gerred05:01:24

to be fair, luminus still beats out all quite a bit, doesn't beat out some node.js frameworks and oddly half as fast as pedestal.

gerred05:01:44

compojure is insanely fast.

shanekilkelly09:01:13

I reckon the objective should be to reflect realistic usage patterns, but it also might be an indication of where to look for potential performance bottlenecks.