A new generation of Tetascale Supercomputer

Kawas is the largest CPU Base HPC System that provides institute-wide service applications in ASIAA, with 16 computing nodes + 2 GPU nodes.

The Kawas system has a total of 2,048 + 192 Cores, and the total capacity is about 1.2PB. With the InfiniBand HDR 200Gbps high-speed network, the overall performance can reach 61.0 TFLOPS.
Kawas has a small number of computing nodes and independent file system space. At the beginning of the design, multiple computing requirements are considered. The computing tasks can be allocated to appropriate physical nodes through the scheduling system. In order to effectively improve system security and avoid the leakage of smart data, The process of logging in to this system requires accessing throught gate server.
This system is a general-purpose computing host that can provide computing services in a variety of fields, including astronomy, physics, mathematics etc. In addition to open-source package software, the system also provides several compilers and MPI libraries for researchers to develop their own calculation programs.

System Specification
Computing node 2 AMD EPYC™ 7763 @ 2.45GHz (64 Cores/CPU) and 512GB main memory
GPU node 2 AMD EPYC™ 7643 @ 2.30GHz (48 Cores/CPU) and 1024GB main memory
8 NVIDIA A100-SXM4 (6,912 FP32 CUDA Cores and 80GB HBM2e/GPU)
Network interface Each node has 1 Mellanox InfiniBand HDR network card, and it also provides MPI parallel computing and parallel file access.
Operational system Rocky Linux x86_64 8.5
Compiler Intel Parallel Studio XE 2020, AMD Optimizing C/C++ and Fortran Compilers 3.2
Queuing system OpenPBS
Connection information kawas.tiara.sinica.edu.tw