HSF / phoenix

"Phoenix", an experiment independent web-based event display for High Energy Physics.
https://hepsoftwarefoundation.org/phoenix/
Apache License 2.0
54 stars 53 forks source link

ATLAS geometry in Safari 15 is corrupted #358

Closed EdwardMoyse closed 1 year ago

EdwardMoyse commented 2 years ago

I'm pretty sure this is an issue with Safari, but it only seems to affect ATLAS geometry.

See screenshot:

Screenshot 2021-09-29 at 14 58 13
9inpachi commented 2 years ago

Did you try toggling performance mode? Anti-aliasing might not be working with Safari but that should affect other experiment too.

On Wed, Sep 29, 2021, 6:01 PM Edward Moyse @.***> wrote:

I'm pretty sure this is an issue with Safari, but it only seems to affect ATLAS geometry.

See screenshot: [image: Screenshot 2021-09-29 at 14 58 13] https://user-images.githubusercontent.com/6764617/135273202-605f1b10-9a9b-4fe9-a468-8652dbd43a95.png

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/HSF/phoenix/issues/358, or unsubscribe https://github.com/notifications/unsubscribe-auth/AIZVY6PWFDKQQAXYTWLK2MLUEME2DANCNFSM5E72MVXQ .

EdwardMoyse commented 2 years ago

No, that doesn't help.

Performance is also TERRIBLE ... 3FPS, and I see the same performance degradation on Chrome too, though there the geometry is rendered okay. It actually make my entire laptop struggle. I tried a few other threejs demos and haven't managed to reproduce the performance degradation there yet, so it's not general.

I can't even open LHCb. :-(

On my iPhone, there are no issues.

I have a Radeon Pro 555 2 GB on my laptop.

I'll keep investigating ... I want to try to open the geometry in other gltf viewers to see if they also struggle with it.

EdwardMoyse commented 2 years ago

In this viewer in Chrome, I see the toroids fine, but performance is terrible:

image

In Safari, same file and same webpage:

image

So it's NOT Phoenix-specific.

EdwardMoyse commented 1 year ago

This was tracked here: https://bugs.webkit.org/show_bug.cgi?id=232367

I believe it is now fixed. Closing.