11-17-2016, 12:50 PM
(This post was last modified: 11-17-2016, 12:52 PM by phillips321.
Edit Reason: i spell like i moonwalk, both shit
)
So then guys,
We're looking at getting a box with 8 GPUs and we're planning on using ESXi on the host and passing through the GPUs to multiple virtual machines.
I've done some quick testing using a single card (AMD R7 240) and i received no performance impact when doing VT-d (yes i know the card is crap but it was used as a PoC)
Has anyone here played with running ESXi and passing through the GPUs to a single VM? My reasons for this is that this will be used in a production environment and the ability to snapshot VMs and quickly increase capability is something we're after.
What i'm after is any real world experience in particularly any info regarding:
We're looking at getting a box with 8 GPUs and we're planning on using ESXi on the host and passing through the GPUs to multiple virtual machines.
I've done some quick testing using a single card (AMD R7 240) and i received no performance impact when doing VT-d (yes i know the card is crap but it was used as a PoC)
Has anyone here played with running ESXi and passing through the GPUs to a single VM? My reasons for this is that this will be used in a production environment and the ability to snapshot VMs and quickly increase capability is something we're after.
What i'm after is any real world experience in particularly any info regarding:
- Did VT-d cause 8GPUs a significant performance impact?
- Is there a ceiling limit on PCI devices in ESXi?