| 9 Mar 2025 |
K900 | And if they decide to split it up into a separate issue, they have gitlab admin tools to do that | 07:19:58 |
eyJhb | Will do :) that shouldn't take too long. | 07:23:19 |
eyJhb | https://gitlab.freedesktop.org/drm/amd/-/issues/3597 | 07:25:22 |
eyJhb | Could also be the same | 07:25:27 |
K900 | That's an APU | 07:26:52 |
eyJhb | https://gitlab.freedesktop.org/drm/amd/-/issues/3445 at least there is a bunch of issues that sounds like they could be VRAM overcommitment | 07:27:10 |
eyJhb | In reply to @k900:0upti.me That's an APU Doesn't that use the same driver? | 07:27:22 |
K900 | Same driver, very different memory behavior | 07:27:32 |
K900 | APUs only have main memory | 07:28:44 |
K900 | dGPUs should only ever use main memory as fallback | 07:28:53 |
woobilicious | okay so Ableton live sees the asio driver yay, now to just figure out how to launch it with the pipewire wrapper | 07:49:13 |
woobilicious | right so proton just ignores $LD_LIBRARY_PATH | 08:21:25 |
eyJhb | K900: doesn't seem to utilize VRAM as much | 09:36:29 |
eyJhb | I'll try to take two new videos. I tried to get htop + amdtop in a video, which was a success. BUT, it's kind of hard when the game is not in frame as well. | 09:36:53 |
K900 | It could also be fragmentation, in theory | 09:37:13 |
eyJhb | I wonder how much of the amdtop is recquired to be shown | 09:42:17 |
eyJhb | Download badv2.webm | 09:50:56 |
eyJhb | K900: ^ is how it looks w/ amdtop + htop + the game runnig | 09:51:08 |
eyJhb | * K900: ^ is how it looks w/ amdtop + htop + the game running | 09:51:14 |
eyJhb | It doesn't seem like it uses all the vram | 09:51:18 |
K900 | Yeah post that on the issue at this point | 09:52:02 |
eyJhb | Yup, on it. Just created a account | 09:52:19 |
eyJhb | Do you have any idea how they prefer videos be uploaded? | 09:52:53 |
K900 | It's 25 megabytes, you can probably just attach it to the issue | 09:53:12 |
K900 | tbh | 09:53:21 |
K900 | Or maybe ask Olivia who is not here | 09:53:39 |
K900 | Who I was about to ping | 09:53:43 |
eyJhb | Olivia Olivia Olivia | 09:53:54 |
eyJhb | Lets see if that works | 09:53:57 |
K900 | She's on US time so probably not | 09:54:17 |