Compute Nodes
The following table compares the compute nodes on Vista to older nodes on Frontera, another HPC resource at TACC. Both systems are supported by the National Science Foundation, and both are intended to enable large-scale computational research in a wide variety of fields. But as mentioned, Vista is designed to be a bridge to the next leadership-class compute facility; therefore, as one might expect, its NVIDIA Grace processors have overall better individual capabilities than those in Frontera. The bigger contrast, though, comes in the sheer number of processors: size is what really what distinguishes Frontera as a leadership-class computer system. The table below emphasizes the numbers and key properties of compute nodes found in each system.
| Frontera1 | Vista2 | |
|---|---|---|
| Compute Node Type | Intel Xeon Platinum 8280 ("Cascade Lake") |
NVIDIA Grace (ARM Neoverse v2) |
| Total Nodes | 8,368 | 256 |
| Cores/Node | 56 (28 cores/socket with 2 sockets) |
144 (72 cores/processor with 2 processors) |
| Hardware Threads/Core3 | 1 | 1 |
| Clock Rate4 | 2.7 GHz Max Turbo: 4.0 GHz |
3.4 GHz |
| Memory | 192 GB DDR4-2933 | 237 GB LPDDR5X |
| Peak Memory Bandwidth | 282 GB/s | 850 GB/s |
| L1 Cache | 32 KB per core | 64 KB per core |
| L2 Cache | 1 MB per core | 1 MB per core |
| L3 Cache | 38.5 MB per socket | 114 MB per processor |
| Local Storage (/tmp) | 144 GB partition on 240 GB SSD |
286 GB partition |
1 Frontera User Guide: System Architecture
2 Vista User Guide: System Architecture
3 As of this writing, hyperthreading is not enabled on Frontera; the ARM processors on Vista are not capable of it.
4 Max Turbo applies when only 1 core is active; the Turbo Boost rate declines if more cores are active and/or AVX instructions are used.
CVW material development is supported by NSF OAC awards 1854828, 2321040, 2323116 (UT Austin) and 2005506 (Indiana University)