NVIDIA 460.67 Linux Driver Brings A Few Fixes, 5.11 Kernel Compatibility

[Templar82]

Actually, open source is exactly why AMD can't implement an equivalent feature. The drivers being open to modification, anyone can comment out any protections.
(Still, open source > closed source, but open also has its drawbacks.)

Exactly.
Why would I want a driver that gimps my hardware?

[mppix]

Actually, open source is exactly why AMD can't implement an equivalent feature. The drivers being open to modification, anyone can comment out any protections.
(Still, open source > closed source, but open also has its drawbacks.)

Nvidia used to be already bad enough with not implementing needed features (wayland).
Are we now starting intentional crippling a feature?

[kylew77]

Anyone know if there will be a binary Nvidia driver for FreeBSD 13.0? I bought a cheap Nvidia GPU for my FreeBSD rig because it is the best supported on that platform but I'm F***ed if they ever decide to stop supporting FreeBSD!

[ix900]

Anyone know if there will be a binary Nvidia driver for FreeBSD 13.0? I bought a cheap Nvidia GPU for my FreeBSD rig because it is the best supported on that platform but I'm F***ed if they ever decide to stop supporting FreeBSD!

You mean the drivers for FreeBSD direct from Nvidia ? because those are here https://www.nvidia.com/Download/Find.aspx?lang=en-us

Just find a guide to install them.

[phoronix_is_awesome]

i'm confused. people on reddit claim that 1337 hackers bypassed the limiters, but this is saying 1337 hackers didn't, rather that nvidia's incompetence left out the limiters in a driver that was released by "accident." so what's true?

From what I hear, Nvidia did *intentionally* or *unintentionally* leak the beta driver. But the vBIOS hack leak is from the manufacturer *MSI* possibly wanting to sell more cards. With Crypto at this price level, greed takes over and all sorts of shit happens.

[d3coder]

Actually, open source is exactly why AMD can't implement an equivalent feature. The drivers being open to modification, anyone can comment out any protections.
(Still, open source > closed source, but open also has its drawbacks.)




Nobody bypassed the protection, the claims were proven to be running other hashing algorithms, that's why the card were reporting more MH/s than expected.
This leaked driver indeed lacked part of the protection, but it was an internal beta (supposedly geared towards testing OpenCL under WSL), so it's pretty buggy and doesn't even work right on all cards.
A blunder on Nvidia's part, but not as monumental as the tech press would have you believe.

This is too bold statement. AMD can implement this in firmware.

[coder]

AMD can implement this in firmware.

Exactly. Not that it'd be as easy to implement, but you'd truly have no recourse if they (or Nvidia) do it in a signed, binary firmware blob.

[coder]

OpenGL 3.0 support...are we going back in time? xD I think you meant OpenCL ;-)

LOL.

Hey, does anyone know if there's anything like Mesa Matrix that tracks different vendors' OpenCL 3.0 feature-support?

[bug77]

This is too bold statement. AMD can implement this in firmware.

Sure they can, but their firmware is not signed/closed so it can be modified at will, too.

[d3coder]

Sure they can, but their firmware is not signed/closed so it can be modified at will, too.

It's closed source and signed too. You can't modify it.