Dear Fellow Amberites,
Following on from the concerns I brought up several weeks ago about NVIDIA deliberately restricting the supply of GeForce cards to companies that sell computers to researchers I wanted to bring your attention to a recent more troubling situation, that I have fought against behind the scenes for many years, involving a change in the end user license agreement that NVIDIA has made in the last few days to it's drivers for GeForce cards.
https://tinyurl.com/ydxfgnjh <
https://tinyurl.com/ydxfgnjh>
Specifically section 2.1.3 which has the new line in bold (my emphasis) below.
-------------------
2.1.3 Limitations.
No Modification or Reverse Engineering. Customer may not modify (except as provided in Section 2.1.2), reverse engineer, decompile, or disassemble the SOFTWARE, nor attempt in any other manner to obtain the source code.
No Separation of Components. The SOFTWARE is licensed as a single product. Its component parts may not be separated for use on more than one computer, nor otherwise used separately from the other parts.
No Sublicensing or Distribution. Customer may not sell, rent, sublicense, distribute or transfer the SOFTWARE; or use the SOFTWARE for public performance or broadcast; or provide commercial hosting services with the SOFTWARE.
No Datacenter Deployment. The SOFTWARE is not licensed for datacenter deployment, except that blockchain processing in a datacenter is permitted.
-------------------
As I am sure many of you will agree this is deeply troubling and does not bode well for the future of cost effective GPU computing. In particular, in my opinion, it speaks volumes about NVIDIA's ultimate intentions. The blockchain exception is particularly Trump like. To me, at least, this implies that in NVIDIA's eyes bitcoin mining is acceptable but science is not. The truth likely being that this is a case of NVIDIA trying to exploit it's monopoly, which unfortunately a number of us CUDA developers unwittingly and pro bono helped NVIDIA build. NVIDIA does not have a monopoly in the cryptocurrency space, hence the exception.
While I am not a lawyer at least for now the EULA appears to be poorly written. It does not define Datacenter or what the term Deployment strictly means. A fact that has been noticed on many forums (e.g.
https://tinyurl.com/ya3qddnf <
https://tinyurl.com/ya3qddnf>). I for one still refer to my clusters being installed in HPC machine rooms. Nevertheless it does not bode well for the future and is likely an omen of what is to come. I would urge each of you who has concerns to contact NVIDIA and make them aware of these.
In the meantime I, and others, are working with AMD to try and complete a port of AMBER, and other scientific and deep learning software, to AMD GPUs to at least try to restore some balance to the force. Anyone who has in-depth experience with GPU programming who would like to help with this effort please do not hesitate to contact me.
All the best
Ross
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Dec 26 2017 - 08:30:02 PST
