I’m only asking this half jokingly, what the fuck?

Unscheduled distributed data duplication

OGC?

I thought we settled that it’s not a good acronym…

https://www.fastprint.co.uk/Assets/User/2004-ogc.jpg

I admit, all distros are a different degree of footguns, I’m saying this as a nix user. lol

What you described is what happaned with arch. The transitioning shouldn’t have happened this way, IMO.

Other distros usually don’t send their users to TTY after an update if they can help it.

On the long term, the situation is the same on linux and windows: you choose the latest driver and live with that given feature set and its bugs.

Rolling distros also only update when you tell them. It is the user who is pulling the trigger on the footgun in both cases.

I’d say the main difference is that arch users are more trigger-happy about being up to date.

Also, I think pacman should at least warn you if the problem is enough to warrant a post on the arch website.

It’s 2025, can we not display a warning message in pacman? Or letting it switch from nvidia-590 to nvidia-legacy?

I’m not an arch user, I admit, I don’t like footguns.

I believe the same SW version is packaged. Nvidia said they’d drop support in the 580 release, but they shifted it to 590 now.

The arch issues are another layer of headache by the maintainers changing the package names and people breaking their systems on update when a non-compatible version is pulled replacing the one with still pascal support in it.

Are they supported longer on the windows driver?

According to the Steam HW survey around 6% of users are still using Pascal (10xx) GPUs. That’s about 8.4 million GPUs losing proprietary driver support. What a waste.

GPU    %
1060    1.86
1050ti  1.43
1070    0.78
1050    0.67
1080    0.5
1080ti  0.38
1070ti  0.24

Fixed: 1050 was noted as 1050ti

Making them open to contributions was the first step, but ok I won’t engage in this petty tribalism.

The topic was about nvidia’s closed source drives.

Valve couldn’t do the same for pascal GPUs. Nobody but nvidia has the reclocking firmware, so even the reverse engineered nouveau NVK drivers are stuck at boot clock speeds.

You can’t expect them to support every single product they’ve ever released forever. And they’re still doing better than AMD in that regard.

If nvidia had the pre-GSP cards’ drivers opensourced at least there would be a chance of maintaining support. But nvidia pulled the plug.

Intel’s and AMD’s drivers in the Mesa project will continue to receive support.

For example, just this week: Phoronix: Linux 6.19’s Significant ~30% Performance Boost For Old AMD Radeon GPUs These are GCN1 GPUs from 13yrs ago.

Make sure to check the compatibility matrix before getting a model not (fully) supported.

Must be an earlier revision. The PM just pinged me to add a fourth and fifth switch for “AI”.

To fulfill the requirements.

I would love to have requirements. (No, vague teams messages still don’t count!)

Oh, look at Mr Many Environments here, he’s too good to test on prod like the rest of us do.

Make it vertical? ¯⁠\⁠_⁠(⁠ツ⁠)⁠_⁠/⁠¯

There are many ethnicities in India and China. Your equivalence breaks down, because then both an uyghur, tibetian and han chinese would be the same.

I get it, if we are talking about outsider’s perspective, I for example can’t tell apart a typical chinese from a typical korean or japanese person. Due to the lack of exposure, my brain only registers them as “asian person”. Maybe the language can give a clue, but that’s debateable.

Still, the broader the brushes we paint with, the more erroneous generalization are.

Why is there an assumption of nationality = ethnicity?

People act the way they act as a reflection of their environment and society, not race.

Would you say you are behaving the same as a mainlander?