Priority Access and Dedicated Storage
TTU faculty members and research grant leaders may purchase priority access to CPU and GPU resources and dedicated storage, in addition to the no-cost resources available to all HPCC account holders. All requests must be submitted via an Operational Agreement in ServiceNow.
Priority CPU/GPU Access:
Provides scheduling priority over the standard fair-share queue on RedRaider partitions, with no limit on job runtime:
| Service | Unit Description | Annual cost/unit |
|---|---|---|
| Quanah (CPU node) |
2x18 core Intel Broadwell, 196GB memory per node, 36 cores total per node |
$1,350 (per node) |
| Nocona (CPU node) |
2x64 core AMD Rome, 512GB memory per node, 128 cores total per node |
$2,500 (per node) |
| Matador (GPU unit) |
NVIDIA Tesla V100, 32GB memory per GPU, 640 Tensor v2 + 5,120 CUDA cores per GPU |
$1,500 (per GPU) |
| Toreador (GPU unit) |
NVIDIA Ampere A100, 40GB memory per GPU, 512 Tensor v3 + 8,192 CUDA cores per GPU |
$1,750 (per GPU) |
Dedicated Storage:
Purchased per terabyte per year, with or without backup. Backup is recommended for important data sets.
| Service | Rate |
|---|---|
| Storage without backup | $40.00 / TB / year |
| Storage with backup | $80.00 / TB / year |
Before you submit
- An appropriate FOP is required. HEAF, NRUF, and TUF funds may not be used for these services.
- Priority access and storage agreements run for one year.
- At expiration, access is disabled, and data must be removed from HPCC storage.
- HPCC storage is not compliant storage. Do not store data subject to PII, PCI, FERPA, HIPAA, or DFARS requirements on HPCC platforms.
Questions
Questions about rates, sizing, or which agreement applies: hpccsupport@ttu.edu
High Performance Computing Center
-
Phone
806.742.4350 -
Email
hpccsupport@ttu.edu