MML Top | Japanese page | About Musashino Media Laboratory
1. Diagram/Abstract
High Performance Computing (HPC) Cluster is under construction for high-speed and massive data computing trial. It is composed of Processing Unit (PU) using slim type desktop PCs, totally ten PUs are connected via two LANs and via KVM to keyboard/monitor/mouse.
|
|
|
| HPC Cluster Diagram | MML Diagram |
PUs are updated in May 2020. PU-#00 / #09 [New] are for operation cluster and PU-#0 / #9 [old] will be for support/backup cluster.
spec. PU-#00 / 09 [New] PU-#0 / 9 [Old] CPU model Intel i5 Intel Core Duo clock 3.2GHz 2.66GHz core/thread 4/4 2/2 Main Memory 8～16GB 4GB Storage SSD HDD GPU nVIDIA GT730 nVIDIA GT730 LAN (NIC) 1Gbps/1Gbps 100Mbps/1Gbps Host OS Windows 10 Window VISTA
2. Application
HPC cluster is planned to use for physical simulation, image rendering and so on. At this time, it is used for Folding@home project, which is dedicated to understanding protein folding using distributed computing technology with PCs all over the world. This project had been started by Stanford University in California USA since 2000, it is fighting COVID-19 virus, cancer, Alzheimer's disease and so on. （ Folding@home Web site )
protein simulation It is easy to join this project just download an application software from project's Web site and install it on PC. Work load is selectable from Light / Medium / Full and also Always working / When idle is selectable. Here are screen shots when Full and Always working are selected.
Operation window Work load graph (GPU) CPU work load Temporally PU-#00 / 09 and other two PCs, totally twelve PUs are running with Always / Full working setting. Musashino Media Laboratory (MML) joins as a team 265080.
Stop (Stand-by) state Working (Simulation running) state Folding@home has a catchphrase "I am One in a Million" and Musashino Media Laboratory may contribute as "I am Tens in a Million".
(c) MML Kaz Tanaka [ Release 2020. 5. 5 ] Please send your comments & request to MML. (Contact)
