Request a future reservation for up to 90 days (in calendar mode)
To reserve TPUs for a specific start time and duration, between 1 and 90 days, request a future reservation in calendar mode. This mode, powered by the Dynamic Workload Scheduler , lets you check TPU availability up to 120 days in advance and request capacity based on your schedule. You can request a future reservation in calendar mode using the Compute Engine API or Google Cloud console. If approved, Compute Engine automatically creates a reservation for your chosen date and time. You can then consume the reservation by provisioning TPUs using the Create Node API or the Queued Resources API . Compute Engine automatically deletes the reservation, and any TPUs that consume it, at the end of your reservation period.
Requesting a future reservation for up to 90 days (in calendar mode) is a good fit for training and experimentation workloads that require precise start times and have a defined duration.
For more information, see About future reservation requests in calendar mode .
View future availability of TPUs
You can view the availability of TPUs 1 to 120 days in advance. For more information, see View resources future availability .
Create a future reservation request in calendar mode
You can request a future reservation in calendar mode using the Compute Engine API or Google Cloud console. For more information, see Create a future reservation request in calendar mode .
Limitations
Requesting a future reservation in calendar mode for TPUs has the following limitations:
- You can't modify, cancel, or delete a future reservation request.
- You must specify a reservation duration with a minimum of 1 day and a maximum of 90 days.
- You can only reserve the following TPU versions in the specified zones:
For the allowed number of TPU chips per request, see Limitations in the Compute Engine documentation.
What's next
- Learn about Cloud TPU reservations
- Share a reservation between projects
- After your reservation start date, consume the reservation
