Create accountLogin

Recent searches

No recent searches

Search options

Only available when logged in.
mastodon.gamedev.place is one of the many independent Mastodon servers you can use to participate in the fediverse.
Mastodon server focused on game development and related topics.

Administered by:

Server stats:

5.1K
active users

mastodon.gamedev.place: AboutProfiles directoryPrivacy policy

Mastodon: AboutGet the appKeyboard shortcutsView source codev4.3.7

Public
thomastc | frozenfractal

Hmm, generating good hashes for use in simplex noise is surprisingly expensive. Has anyone ever tried using the CRC32 instructions that come with SSE4.2 for this? They only take 1 clock cycle, much faster than anything I've come up with.

Public
Steven Dollins

@thomastc Again in GLSL, I like to use the pcg4d hash from jcgt.org/published/0009/03/02/

Perhaps you could adapt it to SSE on the CPU. Or try applying a little bit-swizzling to the CRC32 input or output.

Public
thomastc | frozenfractal

@scdollins Hmm, interesting. I had seen that paper, but was a bit overwhelmed with the number of options. I think I'd want pcg3d, and pass the seed as the third dimension, and then XOR the three outputs together to get the index into the gradient table?

thomastc | frozenfractal

@scdollins Orrr maybe I could interpret the output as a vec2 of floats and normalize it, doing away with the gradient table altogether.

Mar 18, 2025, 07:10 PM·Public
0boosts·0favorites
Public
thomastc | frozenfractal

@scdollins In terms of performance, pcg2d is very promising. Total running time went from 29 ms to 21 ms. I think that would pretty much eliminate the hashing as a bottleneck.

The output still leaves something to be desired though 😅 But who's counting?

A grid of 3x3 simplex noise images. The bottom center one (mine) is completely black.
ExploreLive feeds
About