"The cluster is running great. We were able to complete some production runs and a set of CFD runs that previously took 6.5 days was completed in 14 hours. Thank you for all of your hard work!" -- Mike V., Advanced Clustering Technologies client.
Find out even more reasons to purchase from us
ACT dir is Advanced Clustering's compilation of the best open source software and tools created by Advanced Clustering for high-performance computing. Free and installed on all of our Apex Compute Cluster systems, ACT dir, which got its name because all software is installed and conveniently contained in the /act directory on the head node of your compute cluster, allows our supplied software to be upgraded easily and not conflict with other software you install on your system.
Advanced Clustering's innovative software packages included with the ACT dir installation
ACT dir isn't the only great thing about Advanced Clustering's Apex Compute Clusters. Click here to learn more about included your next cluster purchase from us.
- Breakin — Our hardware stress-test and diagnostic utility.
- Cloner — Our image replication and node installation software.
- Act Utils — Our set of command-line utilities for easy cluster administration and management.
Multiple message passing (MPI) libraries
- MPICH2 for MPI across ethernet networks
- MVAPICH and MVAPICH2 for MPI across InfiniBand networks
- OpenMPI as an alternative to MPICH and MVAPICH supporting both ethernet and InfiniBand networks
- All versions of MPI are compiled with the following compilers:
- GNU Compilers — comes standard with all Linux distributions and our Apex Clusters
- Portland Group Compilers — optional optimizing compiler strongly recommended for our Opteron-based Apex Clusters
- Intel Compiler Suite — optional optimizing compiler strongly recommended for our Intel-based Apex Clusters
- For more information, visit the MPI library websites: MPICH2, MVAPICH, MVAPICH2, and OpenMPI
Easy-to-use environment setup tool
- Pre-configured shell environment setup for both BASH and CSH users
- Automatically makes user-based SSH key pairs on login (if not already created)
- Creates user-based MPD config files with random "secretwords"
- The "modules" environment setup program
- Allows users to switch easily between different compilers and MPI libraries with simple commands
- Preconfigured with modules options appropriate for your custom-configured cluster
- More information on Modules at the Modules project website
The latest InfiniBand software
- Because the Linux distributions may not include the most recent InfiniBand drivers, we always provide them for you as part of the Open Fabrics Enterprise Distribution.
- Providing the latest version of Open Fabrics allows us to support the newest hardware and provides the best performance possible for your investment in InfiniBand hardware.
- More information from the Open Fabrics website.
Your choice of job queuing/scheduler system
- Both Torque and GridEngine are available as part of the ACT dir software package. When you fill out your custom Cluster Order Configuration Form, you can specify which scheduling system you would like installed and configured on your system.
- With either scheduler system install, we provide a full directory of examples on how to submit jobs, which includes single-threaded serial jobs and complex parallel jobs with MPI.
- If you'd like more advanced features or easier setup and configuration than either Torque or GridEngine can provide, we offer the commercial software package Moab. Moab provides a easy-to-use graphical configuration utility and a powerful scheduler with many options.
- More information on the various schedulers: Torque, GridEngine, and Moab.
Monitoring and benchmarking/diagnostic tools
- Every cluster from Advanced Clustering ships with the popular and powerful web-based monitoring system Ganglia. Ganglia provides a current and historical view of your cluster's operation with data collection on a variety of useful metrics, including: memory utilization, processor utilization, disk I/O, network I/O, and more.
- Before each Apex Cluster ships, we run a full series of diagnostic tests with HPC benchmarking tools, such as HPC Challenge, and Intel MPI Benchmarks. These benchmark utilities are installed on your system so you can run them and verify the results yourself.
- Find more information on the Ganglia and HPC Challenge websites.
If you have questions, need assistance, or would like to suggest a feature to any of our software products, please submit a ticket with the details.