#Important resources!

1 messages · Page 1 of 1 (latest)

frank raven
#

The Benchmarking tool for GameMaker!

By @opal harness: https://github.com/DragoniteSpam/GMBenchmark

As each GameMaker update brings new features and optimizations, it's always nice to see which functions are faster than others. This forum serves as a purpose to allow organizing individual sections of optimisations

wind spruce
#

Surely there is no pressure on updating it now. 😆

frank raven
#

You're free to always include things here 🙂

#

And I'll add them

wind spruce
#

Hell no, Dragonite did great work on that! If anything it'll be a PR to his current work.

frank raven
#

👌

visual tusk
#

now ya just got to automate it so that a bot command can compare two different code paths and display the results. Hope you have an enterprise account for automation 😉

frank raven
#

Command line building is free lol

visual tusk
#

Wait, since when

#

I thought that was a enterprise thing

#

Did that change recently or am I just really dumb lol

#

Or was it specific CI/CD pipeline stuff that was behind enterprise

frank raven
#

This was during when we had subscriptions

#

But now you just need an account with GameMaker

#

That's the only literal requirement

#

Not a single penny is needed 🙂

frank raven
#

But that's perfectly fair

#

And fine to be locked behind Enterprise

#

And it is possible to make a bot handle a quick benchmark run too

#

Would have to be basically:
/runbenchmark name_of_branch -> pulls repo branch -> compile + run on linux -> save results to text file -> load in text file and send results

visual tusk
#

shows ya how much attention I pay to things

hybrid lava
#

do we have any tests for how much try/catch slows down code ?

frank raven
#

None at the moment

hybrid lava
#

I'm avoiding using it, as I suspect it affects performance, but I may be wrong, however, I have no idea what test would be best for that

#

iterating 100.00 over array with 100.000, 50.000 and 0 elements, and putting try/catch inside vs. iterating over 100.000 items ?

#

sth like this

opal harness
#

i did a few tests on try-catch a while ago (pre-benchmark tool) and they didnt really slow things down all that much unless you did something ridiculous like log every caught error to the console

#

unclear if that's because try-catch isnt as slow as everyone thinks it is, or if it's because literally everything in gml is just that slow

steady cedar
#

I once tested try-catch as a way to collapse a call-stack and return value. In VM it was fine, but in YYC it was surprisingly slow.

frank raven
#

From what I recall, only catching is slow.

#

If you get no errors, it’s almost no performance hit

#

But I haven’t personally tested that fully, at least in my own game it’s fine

frank raven