These FAQ aim to answer questions regarding the implementation of the flexible GPUs.

What are flexible GPUs?

Up until now, GPUs were provided though specific virtual machine (VM) types which included one or more GPUs (for example, the mv3 type).

3DS OUTSCALE wished to upgrade its offer by allowing to attach or detach GPUs on a regular VM.

Therefore, you are now able to attach or detach a flexible GPU to or from a VM containing the exact number of virtual cores and amount of RAM that your applications require (TINA-type VMs).

How to use flexible GPUs?

All actions on flexible GPUs are done through the OUTSCALE API only.

For more information, see About Flexible GPUs.

When will flexible GPUs be available?

Flexible GPUs are available now through the OUTSCALE API.

What will happen to old VM types possessing GPUs?

New VMs started with a type initially possessing GPUs are converted according to a conversion table, and a flexible GPU is automatically allocated and attached to them. In case you terminate such VM, you must also delete the allocated flexible GPU.

For more information about this behavior change on the existing AWS types, see 3DS OUTSCALE's Offering Refinements FAQ.

How much do Flexible GPUs cost?

Flexible GPUs are charged:

  • When reserved (allocated or attaching states)
  • When in use (attached or detaching states)

Prices vary according to the models of flexible GPUs. For more information, see : https://en.outscale.com/pricing/

The final price of your VM and its associated GPU will not modify your current invoice, which is already calculated based on the following formula: CPUs price + RAM price + GPUs price.

For more information, please do not hesitate to fill in the form to request a call back, or to contact our Technical Support.


AWS™ and Amazon Web Services™ are trademarks of Amazon Technologies, Inc or its affiliates in the United States and/or other countries.