Skip to navigation
Site Primary Navigation:
- About SDSC
- Services
- Support
- Research & Development
- Education & Training
- News & Events
Search The Site:
Condo computing is a shared ownership model in which researchers use equipment purchase funds from grants or other sources to purchase and contribute compute servers (i.e., nodes) to the system. The result is a researcher-owned computing resource of medium to large proportions that provides access to much larger configurations than each participant purchased individually.
Under the TSCC condo program, researchers use equipment purchase funds to buy compute (CPU or GPU) nodes that will be operated as part of the cluster. Participating researchers may then have dedicated use of their purchased nodes, or they may run larger computing jobs by sharing idle nodes owned by other researchers. The main benefit is access to a much larger cluster than would typically be available to a single lab.
Participation in the program runs nominally for five years (the duration of equipment warranty).
The following tables summarize the CPU and GPU nodes that are available for purchase in the year 2022-2023 to participate in the TSCC condo program. Please be aware that costs and configurations are subject to change.
| CPU model | CPU quantity | Memory1 | Estimated Cost |
|---|---|---|---|
| Intel Xeon Gold 6354 (Ice Lake) 18-core 3GHz | 2 | 256 GB (7GB/core) | $11,000 |
| Intel Xeon Platinum 8358 (Ice Lake) 32-core 2.6GHz | 2 | 1 TB (15GB/core) | $20,000 |
| GPU model | GPU memory | GPU quantity | CPU model | CPU quantity | Memory1 | Estimated Cost |
|---|---|---|---|---|---|---|
| NVIDIA Ampere A100 | 40GB | 4 (with NVlink) | Intel Xeon Platinum 8358 32-core 2.6GHz | 2 | 256 GB | $45,000 |
1 Memory size can be increased or decreased based on the requirements. Price will vary accordingly.
2 Other GPU types will be added based on demand. Please contact us if you are looking to purchase a different model of GPU.
The above node costs are estimates; actual prices will vary with market conditions.
A one-time infrastructure fee per node is required to cover the purchase of shared components (racks, network switches, cables, etc.). Researchers also pay a monthly operations fee for each of their purchased nodes. This fee covers labor to maintain the system, utilities, software licenses and other maintenance costs.
| Service | Unit | UC Rate | External Rate |
|---|---|---|---|
| System Setup Fee (covers racks, switches, cables, etc.) | Installation | $2,126 | $3,038 |
| Condo CPU Operations Fee (per purchased node) for UCSD Participants3 | Node/year | $495 | N/A |
| Condo GPU Operations Fee (per purchased node) for UCSD Participants3 | Node/year | $1120 | N/A |
| Condo CPU Operations Fee (per purchased node) for non-UCSD Participants | Node/year | $2736 | $4240 |
| Condo GPU Operations Fee (per purchased node) for non-UCSD Participants | Node/year | $3363 | $5213 |
3Currently for UCSD participants, a portion of the Condo Operations Fee is subsidized by the UCSD Administration and SDSC. Researchers from other UC campuses may be eligible for similar subsidies from their administrations as well.
Typically five years with next business day on-site service. At the end of the period, participants must take possession or surplus their equipment. Obtaining any enclosures, switches, cables, and other ancillary components necessary to make removed nodes operational in a new location is the responsibility of the research group.
Each year, condo cluster participants will receive an allocation of cluster computing time proportional to the capacity of their purchased nodes. For example, a participant that purchases one 36-core node will receive 305,900 SUs. This amount is based on 24x365 usage of 36 cores, allowing for 3% maintenance downtime. These core–hours can be used any time during the year on any of the computing nodes. Unused core–hours by condo participants expire at the end of each year.
Condo jobs that require cores less than or equal to their purchased nodes are guaranteed to start within eight hours of submission and are not subjected to wall time limits. Jobs that extend to other participants’ condo nodes have an eight–hour time limit.
Condo participants may submit "glean" jobs (via the glean queue) to run on idle computing nodes. These jobs are not charged against the submitter's SU balance, but they may be terminated at any time by the scheduler if the nodes where they are running are needed to run higher–priority jobs.
Because the capabilities and purpose of GPU nodes differ significantly from general computing nodes, SUs received for contributed GPU nodes and general computing nodes cannot be interchanged.
Please refer to the TSCC User Guide for further information on node usage, queues and time limits.
Please email TSCC User Support and provide your:
We will contact you to get you a refreshed quote from the vendor and clarify any questions you may have about the TSCC program.