Slashdot Mirror
Nvidia Wants To Prohibit Consumer GPU Use In Datacenters (theregister.co.uk)
The Register reports:
Nvidia has banned the use of its GeForce and Titan gaming graphics cards in data centers -- forcing organizations to fork out for more expensive gear, like its latest Tesla V100 chips. The chip-design giant updated its GeForce and Titan software licensing in the past few days, adding a new clause that reads: "No Datacenter Deployment. The SOFTWARE is not licensed for datacenter deployment, except that blockchain processing in a datacenter is permitted."
Long-time Slashdot reader Xesdeeni has a few questions: Is this really even legal? First, because it changes use of existing hardware, already purchased, by changing software (with potentially required bug fixes) agreements retroactively. Second, because how can a customer (at least in the U.S.) be told they can't use a product in a particular place, unless it's a genuine safety or security concern (i.e. government regulation)!?
Nvidia expects that "working together with our user base on a case-by-case basis, we will be able to resolve any customer concerns," they told CNBC, adding that "those who don't download new drivers won't be held to the new terms."
Long-time Slashdot reader Xesdeeni has a few questions: Is this really even legal? First, because it changes use of existing hardware, already purchased, by changing software (with potentially required bug fixes) agreements retroactively. Second, because how can a customer (at least in the U.S.) be told they can't use a product in a particular place, unless it's a genuine safety or security concern (i.e. government regulation)!?
Nvidia expects that "working together with our user base on a case-by-case basis, we will be able to resolve any customer concerns," they told CNBC, adding that "those who don't download new drivers won't be held to the new terms."
...just because you plaster something in a license doesn't make it automatically law.
"WHQL certified drivers are deployed automatically via Windows Update." -- And if you're running an OS that allows MS to automatically update your server, you deserve everything you get.
09 F9 11 02 9D 74 E3 5B - D8 41 56 C5 63 56 88 C0 45 5F E1 04 22 CA 29 C4 93 3F 95 05 2B 79 2A B2
If I understand it right, you can still do whatever you want with the hardware itself, the restriction is on the drivers (software). That is why it does not apply to those who do not update the drivers. Nor if you use GeForce GPUs on Linux with the Nouveau drivers, but in practice switching to AMD is probably a better alternative.
Unless an organization is already heavily invested in CUDA, they might go with OpenCL instead so they can use AMD consumer stuff instead of Quadros. Even where GeForce versus AMD Vega currently favors GeForce, Quadro prices will make sure that GeForce versus Vega turns that into a win for AMD in terms of investment costs.
In academia, that would also lead to the effect that new developers are more often trained on OpenCL and less on CUDA. That could lead to the sort of long-term win Microsoft Visual Studio had over the Borland development tools.
C - the footgun of programming languages
ATI / AMD wins again!!
As they are open sourcing the ati video drivers in full for Linux.
I don't think this is relevant to GP's argumentation. He (she?) wrote that those who do not update the drivers don't need to agree to the new license. No rooting is necessary here. Legally, it means that the desire of Nvidia to control the use of their hardware can be avoided. At least in the short term.
Practically, the problem will resurface when the current hardware is obsolete/gets unreliable because of age and needs replacing, including drivers for the new hardware. Then the license terms won't be so easily avoided anymore.
Time to switch to AMD. Not only don't they have such clauses, they are actively putting themselves into a situation where a future management cannot easily pull a Nvidia anymore. I mean the open source driver development that gradually replaces the closed source drivers at AMD. Those licenses not be revoked for already released versions.
C - the footgun of programming languages
How is data center defined? I dont have a data center, I have a lab! I dont have a data center, I have a computer room. I dont have a data center I have a mind your own dam business room. Even if in a license I find it hard to see where they can stop the use of a product that has been purchased.
What are the odds that the warranty claims on these cards skyrocketed when they started getting used 24/7 in DC applications and caused this to get thrown in? They know miners won't buy cards that are inefficient for their purposes, and that money train is far too good to throw away, but wealthy corporations and universities? Pay up.
They're leveraging their near-monopoly on GPU-compute and naturally if you have a near monopoly you'll move to exploit that any way you can. I'm thinking this is either (1) a good opportunity for somebody else to come along with new, cheaper compute hardware and make a killing or (2) an anti-trust issue that regulators can take up.
In the end, I will not buy a scumbag's product.
Is if fair for a brick-maker to forbid you from using one for a doorstop?
https://www.youtube.com/c/BrendaEM
Capitalism is funny that way.
AMD and Intel make parallel computing hardware add-ons as well, plus a bunch of little guys. Lesson to be drawn: don't code your stuff in CUDA, use OpenCL so that it is portable to other hardware in case any one vendor gets to big for his britches.
The Halloween documents aren't good enough for you?
The Halloween Documents were 20 years ago, when Microsoft was managed by a completely different group of people. The documents describe a strategy of "Embrace, Extend, Extinguish" that would never work today.
Microsoft is still evil, but their evilness today is very different than it was 20 years ago.
RMS was right. Again.