Hi Fabian,
To add to what Gustaf wrote you should note that the big cloud providers, AWS, Azure etc do not offer GeForce cards. This makes them very expensive in comparison and so if you have facilities to run on premise it can work out a lot cheaper even when you include all potential costs. I have done detailed calculations on this for my day job and we estimate GPU cloud usage, all in, to be between 15 and 25x more expensive than on-premise. The variation is a function of node spec as well as how much data is involved. Remember cloud providers charge you for storage and download of data and those costs can get large quickly.
If you plan on only using compute occasionally and not generating more than a few TB of data then cloud will likely work out cheaper. Similarly if you need a lot of compute for a short period of time cloud resources offer more capability in that arena. However if you plan on running a fairly steady amount of computation, have facilities to host hardware on prem (even better if somebody other than you pays for the power ;-)) then you only need to utilize your nodes about 10% of the time to see significant cost savings over use of cloud resources.
Of course the other difference is that cloud expenditure is often considered opex while hardware is typically capex. So if you have different colored pots of money that may influence the decision.
Hope that helps
All the best
Ross
> On Sep 5, 2019, at 03:25, Fabian Glaser <fabian.glaser.gmail.com> wrote:
>
> Dear Amber experts,
>
> I would like you about your experience about running amber extensive simulations on a cloud (amazon or any other) vs purchasing a new GPU x 4 machine. I have no experience at all with clouds.
>
> Does anybody uses clouds for extensive MD?
> Do you have a cost estimation comparison (per ns x day for example)?
> How difficult it is to deal with implementation and compilation issues on a cloud?
>
> My comparison should be based on purchasing a new GPU X 4 machine, for which I would invest let’s say about 6000$, and comparing it to running on a cloud instead.
>
> Thanks a lot!!
>
> Best,
>
> Fabian
>
> Fabian Glaser PhD
>
> Bioinformatics Knowledge Unit - BKU
> The Lorry I. Lokey Center for Life Sciences and Engineering
> Technion - Israel Institute of Technology, Haifa, Israel
> Web http://bku.technion.ac.il/
> Tel +972 (0) 4 8293701
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Sep 05 2019 - 06:30:02 PDT
