03:47 mangodev[d]: https://cdn.discordapp.com/attachments/1034184951790305330/1420980085908832306/image.png?ex=68d75e55&is=68d60cd5&hm=04c7731c48ece53b9b0da63ac2b5a7ff4298b52d5a7b286c1a26f75264f5361b&
03:47 mangodev[d]: not gonna lie
03:47 mangodev[d]: i briefly considered going back to proprietary because gaming
03:47 mangodev[d]: but the prop driver still kinda feels… not great…
03:47 mangodev[d]: i'm going back to nvk
03:49 sonicadvance1[d]: That's a lot of utilities!
03:50 mangodev[d]: tbf i used `nvidia-all`
03:50 mangodev[d]: also because they have *proper* set-up packages
03:51 mangodev[d]: oh how i missed you, working desktop animations
03:51 mangodev[d]: it'll kinda suck not being able to game for a while more, but things like the work on DCC or the upcoming work on Nova is some hope
03:52 mangodev[d]: although i honestly don't wanna hype up nova too much, it's still effectively a PoC as of now
03:52 mangodev[d]: hardly even a skeleton
03:55 esdrastarsis[d]: mangodev[d]: at least the core driver will boot gsp firmware soon :happy_gears:
04:05 mangodev[d]: my gpu does not have the horsepower to power through 50% less perf in games
04:05 mangodev[d]: but in the desktop, it's fine
04:06 mangodev[d]: esdrastarsis[d]: i hope it handles atomics and stuff better
04:06 mangodev[d]: especially given the whole rust part and all
04:07 mangodev[d]: because nouveau's handling of atomics is… aging…
04:07 mangodev[d]: i use software cursor and it lags far less than hardware cursor
04:07 mangodev[d]: which is the opposite of what it's supposed to be like
09:32 karolherbst[d]: mhenning[d]: doesn't help at all
09:39 karolherbst[d]: but it's pretty obvious that reordering the shader could get rid of ~30 registers
09:39 karolherbst[d]: maybe even 50?
09:43 karolherbst[d]: anyway.. if you have access to the shader, might be fun to investigate and tweak your pass to deal with it
09:43 karolherbst[d]: it's a pretty trivial pattern, just happening a lot
14:10 blisto[d]: @ mhenning renderdoc 1.40 also doesn't crash the game with nvk 🎉
14:15 blisto[d]: https://cdn.discordapp.com/attachments/1034184951790305330/1421138044567687280/image.png?ex=68d7f171&is=68d69ff1&hm=3594766a48f85d121dda2af3fedf9c9eabd892801c5c27e523fee64c57b718fb&
14:15 blisto[d]: https://tenor.com/view/banging-head-bang-gif-12070477
14:16 blisto[d]: I cry
14:23 blisto[d]: <https://github.com/baldurk/renderdoc/blob/a5c3ee4b7051c7a5aacbe73851962516bd2eae68/renderdoc/driver/vulkan/vk_common.cpp#L1389> :TuxThink:
14:28 blisto[d]: Will upload anyway i guess
14:40 marysaka[d]: blisto[d]: huh what version of mesa is that?
14:41 blisto[d]: This commit <https://gitlab.freedesktop.org/mesa/mesa/-/commit/d041640b8877f04e6f68a6e29a4ee1c362d53df3>
14:41 marysaka[d]: or maybe not but we changed something around BDA a bit ago https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37047
14:42 marysaka[d]: hmm
14:56 blisto[d]: https://cdn.discordapp.com/attachments/1034184951790305330/1421148363138662551/image.png?ex=68d7fb0d&is=68d6a98d&hm=381db63b12f50613873cc751014922b7e29e219e89ecf7bf46d5a4674a2a5e6a&
14:56 blisto[d]: Now that i can load in to Call of the Wild: The Angler i spot a visual issue to poke at heh.
14:56 blisto[d]: Small black squared blinking around on the screen
14:58 phomes_[d]: can you check if this helps? https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37207
15:00 blisto[d]: Will try. Gonna try picking it on top of mhennings MR
15:02 blisto[d]: Didn't help sadly
15:18 blisto[d]: Is it expected that `NAK_DEBUG=serial` can cause a hang when trying to load in to the game with it set?
15:18 blisto[d]: [fre sep 26 17:17:04 2025] nouveau 0000:01:00.0: gsp: rc engn:00000001 chid:35 gfid:0 level:2 type:109 scope:1 part:233 fault_addr:0000000000000000 fault_type:00000000
15:18 blisto[d]: [fre sep 26 17:17:04 2025] nouveau 0000:01:00.0: fifo:c00000:0023:0023:[MainThread[34523]] errored - disabling channel
15:19 blisto[d]: I assume that is a hang
15:20 chikuwad[d]: it's the equivalent of an Xid 109
15:20 blisto[d]: k
15:20 chikuwad[d]: iirc
15:20 chikuwad[d]: I could be wrong
15:20 blisto[d]: I see the device lost in the proton log so yea
15:24 chikuwad[d]: ok yeah I'm right
15:24 chikuwad[d]: in this line: `nouveau 0000:01:00.0: gsp: rc engn:00000001 chid:35 gfid:0 level:2 type:109 scope:1 part:233 fault_addr:0000000000000000 fault_type:00000000`
15:25 blisto[d]: lmao what is `trash_memory`
15:25 chikuwad[d]: `type` maps to your regular Xid numbers
15:25 chikuwad[d]: blisto[d]: it writes garbage values to memory before initializing it
15:25 chikuwad[d]: the opposite of zero_memory basically
15:25 blisto[d]: Best config
15:26 chikuwad[d]: chikuwad[d]: `fault_type` 0 is NV_PFAULT_FAULT_TYPE_PDE
15:26 blisto[d]: It's almost a PDF eh?
15:26 chikuwad[d]: https://github.com/NVIDIA/open-gpu-doc/blob/87ba53e0c385285a3aa304b864dccb975a9a0dd4/manuals/turing/tu104/dev_fault.ref.txt#L23
15:27 blisto[d]: Is that a document telling which dev is at fault?
15:27 blisto[d]: What is PDE?
15:28 blisto[d]: The last one was a serious question :frog_pregnant:
15:28 chikuwad[d]: page directory entry if I'm not mistaken
15:28 blisto[d]: k
15:29 chikuwad[d]: basically a pointer to a page table
15:33 blisto[d]: The game have a hang issue on nvidia (maybe a vkd3d-proton issue, but don't know) so was curious if i could trigger the same on NVK. Sadly the issue depends on fairly high ish FPS to trigger (~160fps ish on nvidia) and NVK only have around 60-70.
15:33 blisto[d]: Not that i would expect it to hit the same, so was just a experiment