Building Unorthodox Deep Learning GPU Machines

Re:Building Unorthodox Deep Learning GPU Machines by kookamamie on Wednesday February 28, 2024 @12:13AM (#39534258)

Do note that, due to NVidia's absurd EULAs, you cannot run RTX 3090 in a data-center for compute purposes (https://www.nvidia.com/content/DriverDownloads/licence.php?l...).

Should you care about the EULA? Not really, until you're a business of any significant scale.

[ Parent | on HN ]

Re:Building Unorthodox Deep Learning GPU Machines by echelon on Wednesday February 28, 2024 @04:18AM (#39535742)

If you update your drivers. You could use an older license.

[ Parent | on HN ]

Re:Building Unorthodox Deep Learning GPU Machines by hkgirjenk on Wednesday February 28, 2024 @05:08AM (#39536051)

Nvidia is known to retaliate (delaying future shipments) if you do stuff like this.

[ Parent | on HN ]

Re:Building Unorthodox Deep Learning GPU Machines by PrayagBhakar on Wednesday February 28, 2024 @01:21PM (#39541785)

I run a basement compute server[^1], what’s Nvidia gonna do? Not let me buy their hella expensive H100s? At least now I get to learn ML skills without my failed experiments exponentially scaling on the cloud.

[^1]: https://prayag.bhakar.org/apollo-ai-compute-cluster-for-the-...

[ Parent | on HN ]