Linux Compute Intensive Systems
OIT RC maintains a number of computational systems for academic research use. These systems are typically used for long-running, computationally intensive, and data-intensive processes, as well as processes with specialized requirements, such as a parallel computational environments supporting standard MPI libraries, GP-GPU computing, or very large memory requirements. These servers have access to a large array of research-related software for non-parallel and parallel scientific computing tasks.
Linux Compute Intensive Systems Specifications:
Circe (circe.rc.pdx.edu)
Dell PowerEdge R720
2 x Intel(R) Xeon(R) CPU E5-2665 with 16 cores @ 2.40GHz
192 GB RAM
10 TB local scratch (/disk/scratch/)
Best used for smaller jobs
Hecate (hecate.rc.pdx.edu)
Dell PowerEdge R720
2 x Intel(R) Xeon(R) CPU E5-2690 with 16 cores @ 2.90GHz
768 GB RAM
10 TB local scratch (/disk/scratch/)
Best used for high RAM jobs, such as more than 100 GB of RAM being used.
Agamede (agamede.rc.pdx.edu)
Dell PowerEdge R730
2 x Intel(R) Xeon(R) CPU E5-2695 v3 with 28 cores @ 2.30GHz
256 GB RAM
8 TB local scratch (/disk/scratch/)
Best used for high multi-threaded jobs.
Tuning Your Software
The optimal number of threads to use is entirely dependent on the program being run and may take a bit of trial and error. More threads does not usually result in faster runs and can actually be detrimental. Start small and increase the concurrency of your program. Benchmarking will help you discover the optimal settings for your program.