10x1080ti for rent for deep learning, rendering...

[vprelovac]

Hi everyone

I am curious if anyone would be interested in renting a 10x 1080ti server (256 GB RAM, 7 TB SSD, 2x Intel Xeon E5-2698 v3, 1Gbps bandwidth)?

Rented as baremetal online, Windows and Ubuntu available.

Offered at $125/day, in DC in Sunnyvale, CA. 10% discount available for STH members.

On premise lease also possible (SF Bay Area only, $999 per week).

Happy to answer any questions.

Vladimir
email: sales@teclis.com

 

[K D]

I hope you have @Patrick 's approval for this commercial posting.

 

[Patrick]

I hope you have @Patrick 's approval for this commercial posting.

Approved. I have been working with him on this for some time. Certainly, an interesting service if you want to do deep learning training or rendering.

 

[Revrnd]

@Vladimir, if you keep an eye on RNDR, you might be able to use some of that spare compute power to go toward distributed 3D Rendering.

OTOY have a Beta they're going to be testing soon (sign-up closed unfortunately) but upon successful completion of the Beta, you might be able to get an idea of the returns for RNDR.

 

[vprelovac]

@Vladimir, if you keep an eye on RNDR, you might be able to use some of that spare compute power to go toward distributed 3D Rendering.

OTOY have a Beta they're going to be testing soon (sign-up closed unfortunately) but upon successful completion of the Beta, you might be able to get an idea of the returns for RNDR.

Thanks for the tip. I have been watching that space and it is great that GPU compute found its way in. Another model that can be useful for some people here is what VectorDash.com is doing.

 

[ideabox]

Thanks for the tip. I have been watching that space and it is great that GPU compute found its way in. Another model that can be useful for some people here is what VectorDash.com is doing.

Except Vectordash took too long, and Vast.AI is up and running in the space.

 

[LukeP]

how is any of those different to every other gpu cloud instance by google, aws, azure, paperspace, .... ?

unlimited gpu has been available for years.

its only a matter of tflops per $. all the rest is fluff.

 

[ideabox]

You answered your own question:

its only a matter of tflops per $.