Although the evolution of the Open Source software used to create an HPC cluster has continued steadily, creating a complete functioning cluster requires specific knowledge and expertise. Integrating the disparate software components required to control the hardware in order to provide a usable resource with integrated management, system administration, resource allocation and reporting features is no simple task.
For the past 12 years SICORP has supported HPC cluster administrators by being involved in the continuing evolution of the HPC software stack that provides for integrated supercomputing functionality.
Operating System & Cluster O/S Provisioning:
- Red Hat® - Red Hat Enterprise Linux (RHEL) 64 bit operating systems
- CentOS - Open Source 64 bit operating wystem based on RHEL
- Bright Computing® - Comprehensive scalable cluster installation and management software
- StackIQ® - ROCKS+ - Commercially supported versions of ROCKS
- Perceus - Open Source disk-less cluster O/S and node provisioning
Cluster Resource Management & Cluster System Administration:
- Adaptive Computing® - Intelligent management middleware that provides simple web-based job management, graphical cluster administration and management reporting tools
- SGE- Sun Grid Engine Open Source scheduling and reporting tools
- Maui/ Torque - Open Source resource management, scheduling and reporting tools
- LSF® -Workload manager for demanding, distributed and mission-critical HPC environments
- SLURM - Simple Linux Utility for Resource Management
Cluster Hardware Management & Monitoring:
- Nagios - Open Source network monitoring tools
- Ganglia - HPC Admin, Login and Compute node monitoring
- IPMI & BMC control - ConMan, PowerMan
- LM Sensors- Base hardware monitoring
Compilers, Code Development and De-bugging Tools:
- GCC - Fortran, C, C++ compilers and debugging tools
- Intel® - Fortran, C, C++ compilers and debugging tools
- PGI® - Fortran, Fortran90, C, C++ and CUDA-Fortran compilers and debugging tools
- TotalView® - High level code debugger and clustered code monitoring debugger
- OpenCL - Open Source compiler and development tools aimed at GPGPU computing
- CUDA® - Open Source development tools aimed at NVIDIA Tesla GPGPU computing