Training
Aspen Systems provides on-site and off-site training services for all aspects of your HPC needs. Our professional staff can help ensure your administrators and end-users have the information and skills they need to best utilize their new or existing HPC resources. Whether you need formal classroom training for your staff, informal training on-site directly with your users and administrators, or entirely custom curricula designed to your specific requirements, Aspen can help!
Unlike some other vendors, Aspen offers informal, free flow, "one on one" training sessions on site with your administrators and end-users, often as part of an on-site installation visit. Many customers find this method extremely effective, and enjoy the flexibility of having our training resource on site with their users and working on their specific problem sets with them in their own environment.
If a more traditional or structured training environment is desired, Aspen can provide training facilities in your area or at Aspen to accommodate your needs. To be cost effective for larger class sizes, Aspen recommends that we send an instructor to you and utilize training facilities at your organization or in your area.
Aspen provides basic training packages for cluster users and administrators which are tailored to be most effective on an Aspen configured cluster. All of our standard training presupposes some degree of familiarity with Linux commands, and utilizes that knowledge as a base to introduce more advanced concepts. We can also offer custom training sessions for your specific needs as well.
Aspen Training Packages
Aspen Cluster User Training covers a variety of standard topics many new cluster users have difficulty with, and is intended to take the user from their first login on the new cluster to actually running their codes with real world problem sets. Aspen will need a list of your users applications in advance in order to prepare properly for your training session, and remote access to your cluster from the training location is highly recommended.
- Cluster Access: Completion of this training module will allow the end user to connect to their cluster via secure shell or the organizations selected methods, such as VNC, ABC, X Windows clients, or legacy "R" commands.
- Cluster Familiarization: In this section, the cluster users learn more about their new systems. Training goals are to successfully identify system provided architecture(s), memory density, Interconnects, discrete node type and quantity, and the usage of each.
- Cluster Environment: This training teaches your users how to use environment modules or other methodologies to select between installed compilers, MPIs, and utilities on the delivered system.
- User Codes: One of the most daunting tasks new users face is how to build their codes. This module trains the end users to modify Makefiles or application installation scripts to build correctly with their selected cluster environment. Additional utilities, library paths, and other variables specific to the users particular code are outlined as well.
- Code Execution: Almost every code can be ran interactively for troubleshooting and verification, or via the installed scheduler or vendor UI for routine operation. The goal of this module is to show your users first how to do manual runs (for code verification or regression testing) and then to help them write and execute scheduler submission scripts or utilize the software UI as needed to submit jobs for code execution. Successful completion of this training ensures that the end user can successfully execute codes on their delivered cluster.
- User Data: This module is designed to train the users where to store data on their cluster, and how to access other post processing utilities as required by their code(s). If the cluster is configured for external data access, this module will train the users how to mount their file systems on their workstations or retrieve the data to other systems in the organization as necessary for further processing.
