Nvidia has banned the use of its consumer-grade GPUs in data centers, the Register reported. This decision forces organizations to pay for the more expensive chips such as the Tesla V100. The GPU company updated the license agreements of its Titan and GeForce software to reflect this change.
“No Datacenter Deployment,” the updated agreement reads. “The SOFTWARE is not licensed for datacenter deployment, except that blockchain processing in a datacenter is permitted.”
Prior to this change, it was possible for data centers to use several Titan or GeForce GPUs in servers in order to power things such as A.I. research, machine learning, and simulations and analytics. Technically speaking, the EULA only covers the drivers, but without Nvidia’s drivers, you can’t access the full power of the hardware, so the Titan and GeForce will do you little good.
This means that data centers and hardware manufacturers building workstations for those centers will have to fork out a lot more money for GPUs. The gamer-focused GeForce GTX 1080 Ti only costs about $700. It’s not cheap, but it is much more affordable than the $9,000 Tesla V100.
In a statement to the Register, Nvidia said that it updated the license agreement to reflect the fact that it does not believe its consumer-grade cards are meant for data center work.
“GeForce and TITAN GPUs were never designed for data center deployments with the complex hardware, software, and thermal requirements for 24×7 operation, where there are often multi-stack racks,” a spokesperson for Nvidia said.
However, some researchers have said that they do not need the extra power offered by the Tesla V100 stating. One research expressed frustration that Nvidia would resort to legal threats against researchers for using consumer-grade cards.
“Much of this data does not actually need to double-point precision of the Tesla cards,” the researcher told the Register. “This is a shocking sneaky way to suddenly introduce limitations to their product. Most places would be happy to acknowledge that using a consumer product in a server may invalidate the warranty, but to limit in such a way that it would leave researchers open to possible legal threats is frankly disgusting.”
EULAs are often difficult to enforce, though it is possible that Nvidia could take organizations to court over the violation of the agreement.
- Bosch, Daimler team up to deploy autonomous Mercedes-Benz S-Classes in San Jose
- Nvidia’s RTX 2000 GPUs look like A.I. hardware dressed up for gamers
- Lensless cameras could turn windows into sensors, even pointed the ‘wrong’ way
- Teaching machines to see illusions may help computer vision get smarter
- JackRabbot 2 is Stanford’s friendly new campus-roaming social robot