Choosing a GPU server

Started by glawar, Jul 05, 2022, 11:56 PM

Previous topic - Next topic

glawarTopic starter

Good afternoon,
tell me, are servers with A5000 cards suitable for complex calculations, in particular, for modeling thermodynamic processes?
Just that now there is a need for computing power. Of course, we will not buy a server for this, we will rent it.
They are not cheap. Models are purely computational.
  •  

evejones

We rent servers in the Netherlands from Hostkey.
GPU with RTX A5000 cards and an IPMI port. They launched a web control panel last year. It's very convenient to install the OS and other software using ready-made and custom ISO images. There are no problems with the server for processing 3D parts processing models in Solid.
Nothing falls, the channels are good.
  •  

adm.1.n

If you really going to use this server just for thermodynamics processes then yes, i guess it's gonna work well. By the way, the rent of this kinda servers became a bit cheaper these days.
  •  

Edison Duncan

I did not expect to read on this forum the question, will the A5000 have enough power to simulate thermodynamic processes? Many don't even know what it is. But all this is a subject. My father is just doing this, and you know, at their university, in the laboratory, they don't have a system with a strong GPU, but they have a strong processor.
If you need a server with a strong map, then I recommend paying attention to:
gpu-mart.com/rtx-a5000-hosting
linode.com
paperspace.com/pricing
  •  

Yulia

I suggest, based on the question - take the most powerful, expensive and beautiful. The calculator and VirtualBox will start on a large amount of modern budget hardware
  •  

dasiydan88

According to the results of testing, the A5000 wins in terms of price and quality.
The best result in OpenCL Compute Score is slightly inferior to the more expensive A40 and A100 in CUDA Compute Score and is suitable for working with graphics. Second place in AI-benchmark after A100. The leader in the V-Ray rendering speed test, the leader in the transcoding test. Supports VDI. Unconditionally our option, if compared with the price table.

The A2000 is five times cheaper than the A5000, while demonstrating acceptable benchmark results for the base model. Does not support VDI, but is suitable for working with graphics and AI tasks.

The A4000 is "average" in performance between A2000 and A5000, does not support VDI, but otherwise withstands criticism in terms of price and benchmark results.
A100, as I have already written, is the undisputed leader for working with artificial intelligence, model training, inference, data analysis and complex calculations. It is optimal for remote desktop infrastructure.

The remaining GPUs showed less results when comparing benchmarks and prices.

There are four video cards on the podium of the winners (which, by the way, can already be ordered on the site). Our formula corresponds to RTX A2000, RTX A4000, RTX A5000 and A100.

From inexpensive web servers with one GPU to clusters with multiple graphics cards on board. If you need a "strong" server for rendering, add A2000 to it — it will do the job for five and will not "eat" the budget.
And for ambitious tasks with complex calculations, AI, large VDI projects, there is a server with eight A100. There is already a ready-made config. Despite our strict selection, we are ready to provide the client with any NVIDIA card (except desktop RTX 3080 and 3090, of course).

The selected cards are in stock, which means that you will receive a custom web server with them within five days. If an already assembled server with GPU is suitable, it will be ready for operation in 4-50 minutes.
  •