PC-based Distributed Memory Multi-Processors Cluster Computing


Cluster Computing System: Two rack cabinets on the left are the old HCSERVER, and the rack cabinet on the right is OP64SERVER
ES69 server, external raid disk and monitor
Offline Backup and file server system for OP64SERVER, HCSERVER and ES69 systems

“Cluster Computing System” is a type of parallel processing hardware facility using distributed memory and pc- based CPUs under efficient network communication configuration to accomplish the tasks which demanding either under Single Instruction Multiple Dataset(SIMD) or Multiple Instruction Multiple Dataset(MIMD) type of highly parallelized computer jobs. The system can be operated even when part of the computing node is not functional or under maintenance.

For efficient and highly parallelized cluster computing system, stable operation system; application code modification and parallelization strategy are the necessary ingredients

OP64SERVER system has 64 bit, 2CPU 8Gb memory; HCSERVER system has 32 bit, 6 single CPU core; 24 dual- codes. Each node has 2Gb memory; ES69 server system is a 64 bit, 8 way, shared memory parallel machine with 64Gb memory equipped with 1 TB disk storage and backup hard drive. All systems are using Linux operating system.


TSMIP-CWB strong motion network and mechanisms of 331 Hualien earthquake A U-D component of 331 Hualien earthquake record collected by CWB strong motion network
A 3- D perspective view of earthquake seismicity, topography and velocity model in northern Taiwan Partition of composite grid system in distributed cluster computing
A fixed- time snapshot of strong motion wavefield simulation in Taipei basin Basin amplification from finite- difference numerical modeling