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.

Sign up for GitHub

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

Jump to bottom

OPF example: big perf regression between 0.8.1 and tip on AMD #3157

Closed
teh-cmc opened this issue Aug 30, 2023 · 6 comments
Closed

OPF example: big perf regression between 0.8.1 and tip on AMD #3157

teh-cmc opened this issue Aug 30, 2023 · 6 comments
Labels
🚀 performance Optimization, memory use, etc 🦟 regression A thing that used to work in an earlier release

Comments

@teh-cmc
Copy link
Member

teh-cmc commented Aug 30, 2023

main:
image

0.8.1:
image
@teh-cmc teh-cmc added 🚀 performance Optimization, memory use, etc 🦟 regression A thing that used to work in an earlier release labels Aug 30, 2023
@jleibs
Copy link
Member

jleibs commented Aug 30, 2023

Two more datapoints:

Intel + Wayland

image
Something different (but also not good) going on here.

Nvidia + XWayland

image

@teh-cmc
Copy link
Member Author

teh-cmc commented Aug 30, 2023 •
edited
Loading

The regression was introduced in effb72b.

Just bisected into it, haven't looked at the underlying reason yet.

Also I pretty vividly remember reviewing and running that PR on my laptop without any issue back in Stockholm... please don't be a dedicated-amd-gpu-on-linux-through-mesa kinda thing 🙏😭

@teh-cmc
Copy link
Member Author

teh-cmc commented Aug 31, 2023

I cannot reproduce the issue on my laptop: both the embedded intel GPU and the discrete nvidia card give about 55ms per frame.

My laptop has the exact same OS/software setup as my desktop so... I guess this is an AMD thing 😒

@emilk
Copy link
Member

emilk commented Aug 31, 2023 •
edited
Loading

The regression diff should only touch CPU code; the GPU should see the exact same data and calls o.O - some weird timing thing? Or some difference slipped in somehow

@emilk emilk changed the title OPF example: big perf regression between 0.8.1 and tip OPF example: big perf regression between 0.8.1 and tip on AMD Sep 20, 2023
@emilk
Copy link
Member

emilk commented Oct 9, 2023

For reference, on my MacBook Pro the performance has improved by a factor of 3x from 0.8.2 to 0.9: #1136 (comment)

@emilk emilk removed this from the Triage milestone Nov 6, 2023
@teh-cmc
Copy link
Member Author

teh-cmc commented Jul 9, 2024

Mesa 🤷‍♂️

@teh-cmc teh-cmc closed this as not planned Won't fix, can't repro, duplicate, stale Jul 9, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
🚀 performance Optimization, memory use, etc 🦟 regression A thing that used to work in an earlier release
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
3 participants
@emilk @teh-cmc @jleibs