Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fresh screenshots? #6

Open
kvark opened this issue Jul 25, 2020 · 5 comments
Open

Fresh screenshots? #6

kvark opened this issue Jul 25, 2020 · 5 comments

Comments

@kvark
Copy link

kvark commented Jul 25, 2020

I can see that Terra is now ported to wgpu, based on Cargo manifest. However, the README still has this:

Terra is a large scale terrain rendering library built on top of gfx pre-ll.

It's also unclear if the screenshot is from wgpu updated renderer. I guess it's not.
We'd be happy to feature your work on https://wgpu.rs/#showcase, so we'd need an to date screenshot.

I tried to run it locally, and wasn't able to:

thread 'main' panicked at 'called Option::unwrap() on a None value', bin/preview.rs:80:10

@fintelia
Copy link
Owner

Thanks for the reminder to update the screenshot and readme! I'd been meaning to do it but hadn't gotten around to it...

I'm not quite sure what is causing the error you are seeing. Based on the line number it seems to be that it is failing to create an adapter. I just pushed 8431a34 which changes from asking for a Vulkan backend to any primary backend, so perhaps that may fix things?

@kvark
Copy link
Author

kvark commented Jul 26, 2020

That's better, thank you! Looks like we are hitting an MSL translation issue, internally, now:

There was a compiler error: double types are not supported in buffers in MSL.

@fintelia
Copy link
Owner

Hm, it could be tricky for Terra to handle hardware that doesn't support doubles. Do you happen to know where Metal does/doesn't allow them?

@kvark
Copy link
Author

kvark commented Jul 27, 2020

According to MSL 2.1 spec, Metal doesn't support Double at all...

@fintelia
Copy link
Owner

fintelia commented Aug 1, 2020

I looked more into dealing with this and there's now a branch (no-double) that switches everything to floats. That works for now, but will have precision issues once fractal refinement can generate features smaller than a few meters. I also tried fixed precision arithmetic, but while I was able to get something running, it seemed that it was going to be a big hassle going forward not to mention slower and less precise as well.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants