Available compute resources
As mentioned in the Service overview, Taiwania2 (HPC CLI) shares 252 computing nodes and 2,016 NVIDIA® Tesla® V100 GPUs with some TWSC computing services. Taiwania 2 has excellent performance of 9 PFLOPS (9 Peta floating-point operations per second), supports up to 300 GB/s transfer rate between GPUs with NVLINK, also provides a high-bandwidth, low-latency parallel computing network.
The specifications of a single node are as follows:
| Item | Description | Quantity |
|---|---|---|
| CPU | Intel Xeon Gold 6154 18 Cores 3.0GHz | 2 |
| Memory | 32GB DDR4-2666 RDIMM | 24 |
| System disk | 2.5” 240GB SATAIII | 2 (RAID 1) |
| HDD (Hard disk drive) | 4TB NVMe | 1 |
| GPU | NVIDIA® Tesla® V100 SXM2 | 8 |
| Network card | Mellanox InfiniBand EDR 100Gb | 4 |