When you deploy an interruptible instance, you're setting a bid price for the compute price of the instance that you're looking to deploy.
Storage prices are billed at the standard rate, and are added atop the compute price that you bid. If you bid a price for $0.35/hr with 200GB of storage at $0.00005/GB/hr, then your total cost will be $0.36/hr when the workload is running and $0.01/hr when outbid.
The bid price you set is per-virtual machine, not per GPU. We prioritize workloads for max revenue, not the per-unit price of resources. For instance, if there are these bids for a 3x A6000 machine:
- $0.25/hr for 1x A6000 ($0.25/hr/A6000)
- $0.27/hr for 3x A6000 ($0.09/hr/A6000)
Then the 3x A6000 will win out, even though it has a lower per-unit price because the overall revenue generated is greater.
However, if there are these bids for a 3x A6000 machine:
- $0.25/hr for 1x A6000
- $0.03/hr for 1x A6000
- $0.27/hr for 3x A6000
Then the 1st and 2nd A6000s will win out because the combined revenue of those two VMs ($0.28/hr) is greater than the revenue of the alternative option ($0.27/hr).
Note that these are for illustrative purposes only. Each host has a minimum bid price set per GPU. These are typically around 50% of the on-demand price.
Because we optimize for aggregate revenue and not per-unit revenue, we expect a few sweet spots:
- High CPU/RAM allocations: deploying more resources isn't more expensive, as you just need to beat out everyone else
- Very low pricing on less-used GPUs: with fewer bidders, P4s, T4s, V100s, etc. will be much lower priced on an adjusted cost-to-performance basis than GeForce alternatives
- Low bids on single GPU VMs might win out
To deploy an interruptible instance, set the following fields in your API request:
price: [your bid]
Interruptible instances will appear within your dashboard like the image above.
Interruptible instances are only available on GPU instances, and storage costs are billed at the standard storage price set by the host. There is no "interruptible storage" tier, so to speak.