Save 10%Pay yearly on any plan. 6% half-yearly, 3% quarterly. Help centreWhatsAppTelegram TeamsOpen a ticket

GPU RDP or GPU Dedicated Server: Which Do You Need?

Two ways to rent a graphics card: a shared GPU RDP or a whole GPU server. Five questions decide between them, with a table of jobs and the right plan.

3 min readBy Shailabh

Reviewed by the AmazingRDP technical team

GPU RDP or GPU dedicated server

A Windows desktop with a graphics card, shared, at a fraction of a whole server's price. See GPU RDP from $29.99

There are two ways to rent a graphics card from us. A GPU RDP is a remote Windows desktop on a server whose card is shared among its users. A GPU dedicated server is a whole machine whose card is yours alone. They differ several-fold in price, and the right one depends on five questions. Answer them in order; the first "yes" in the right-hand direction settles it.

The two at a glance

GPU RDPGPU dedicated server
You getA user account on a shared serverThe whole physical server
Graphics cardShared with the other usersYours alone
Administrator accessNoYes
Install software and driversNo; common programs are installedYes, anything
SystemWindowsWindows (evaluation) or Linux
LocationEuropeEurope
Ready inMinutesHours; some builds longer
PriceLowSeveral times higher
Change of planBy ticketA new order

Question 1: for how long will the card be busy?

A shared card is for work in bursts: an hour of editing, an export, a look at a 3D model. Jobs that occupy the card for many hours, or all day, are not allowed on the shared plan, because they would take it away from everyone else. Long renders, all-day streaming and round-the-clock processing need a dedicated server.

Question 2: do you need to install anything?

On the GPU RDP you use what is installed and cannot install programs that need administrator rights, nor drivers. If your work depends on a particular application, plug-in, driver version or a framework such as CUDA libraries, you need the dedicated server.

Question 3: is it an Android emulator, or a bot?

Emulators are allowed only on GPU dedicated servers. Bots are not allowed on the GPU RDP either; they belong on a Botting RDP or a dedicated server.

Question 4: do you need Linux?

The GPU RDP is a Windows desktop. Machine-learning tools mostly live on Linux, which means a dedicated server, where you install the system and the drivers yourself: NVIDIA drivers on Ubuntu.

Question 5: does the card's full power and memory matter?

On a shared card, other users' work takes part of its speed and video memory at any moment. If you must be able to count on all of it, for a render deadline or a model that only just fits, take the dedicated server, and compare the card model with your software's requirements: choosing a GPU dedicated server.

Jobs and the right plan

JobPlan
Editing and exporting a video now and thenGPU RDP
Viewing or lightly editing 3D modelsGPU RDP
Web applications that need graphics accelerationGPU RDP
Trying whether your workflow works remotely at allGPU RDP first
Rendering for hoursGPU dedicated server; for processor rendering, Rendering RDP
Streaming around the clockGPU dedicated server, or the Streaming RDP plans
Android emulatorsGPU dedicated server
Machine learningGPU dedicated server, with enough video memory
Converting a video libraryEncoding RDP, or a dedicated server
Cryptocurrency miningNot allowed on any plan

What neither is

Neither is a gaming service. Remote Desktop limits the frame rate and adds delay; slow-paced titles may be playable, fast ones are not. And a card in a datacentre does not improve the picture beyond what your own connection can carry: on a weak line, lower the resolution. Practical settings are in how to use a GPU RDP.

A sensible path

  1. If none of the five questions pointed to dedicated, start with the GPU RDP for a month. The refund period is 7 days with used days deducted, and a trial of a shared plan can be requested.
  2. If you hit a limit (a program you cannot install, a job that runs too long), you will know exactly which dedicated build you need.
  3. Moving means copying your files; the two are different machines.

The background on what graphics cards do beyond games is in eight jobs for a GPU; the product itself in the GPU RDP guide. Plans and current prices: GPU RDP and GPU dedicated servers.

Questions

What is the difference between a GPU RDP and a GPU dedicated server?

A GPU RDP is a user account on a server whose graphics card is shared. A GPU dedicated server is a whole machine with a card only you use and full administrator access.

Which is cheaper?

The GPU RDP, by a wide margin, because the hardware is shared.

Can I render for many hours on a GPU RDP?

No. Long jobs that occupy the shared card are not allowed; use a GPU dedicated server.

Can I install my own drivers?

Only on a GPU dedicated server.