Draft

This documentation is in draft and under active review. Figures for the Fall 2026 GPU reservation system are not yet published, and some pages describe behavior that has not been verified against a primary source. Confirm any figure with datahub@ucsd.edu before relying on it.

GPU Access

This section covers GPU access on DSMLP under the Fall 2026 reservation model: GPU classes, reservations, Service Units, quotas and availability, and the causes of a session ending.

Page Covers
GPU Classes GPU class sizes, choosing and requesting a class, preset classes on Datahub, a missing or misspelled class label, and confirming the allocation.
Reservations Reservation types, booking ahead and launching without a booking, the claim window, length caps, best-effort reservations, team mode, and routes to more time.
Service Units & Budgets What spends Service Units, hourly rates, budget sizes and windows, the cancellation penalty, and the routes available when a budget runs out.
Quotas, Cohorts & Availability GPU quotas, borrowing beyond quota, cohorts, the symptoms of a full cluster, the status page, and requesting a quota increase.
What Ends a Session Idle culling, the end of a reservation window, overstay, preemption, and maintenance closures.

← Documentation index