The SystemX architecture streamlines the amount of time scientists need for research. Some problems may take literally months of computer time. SystemX may reduce that to a few hours. In addition, there are classes of problems that may only ultimately be solved using the cluster capabilities of SystemX.
Solving Different Questions, Differently
SystemX encourages researchers to ask questions about which they may only have dreamed. From simulating hundreds of biochemical pathways in the living cell, to evaluating billions of molecules for inclusion in a new drug, terascale computing means problem solving at unprecedented speeds and accuracies.
Corporate researchers have rarely had access to computational resources of this class. Even when they may have been able to use a super-computer, they have seldom had the support and training necessary to port and optimize code for the systems that they have used. TeraDisc changes all of that. From its dedicated facilities on the same campus as SystemX, connected to the cluster by high-speed fiber optic networks, a staff of engineers and graduate students helps member-researchers fine tune programs for use on the cluster. In the case of high-security research projects, TeraDisc even has the ability to completely isolate the cluster, load an encrypted research dataset, run the code, and return the output in encrypted form before scrubbing the cluster’s multi-terabyte mass storage systems and placing the cluster back on line.
Problem characteristics that are well-suited to TeraDisc's approaches include:
- Problems that feature multiple tasks that can be replicated across multiple processors;
- Problems that require large memory;
- Problems with many repetitive serial tasks;
- Problems that must process data from multiple sources in multiple ways;