Skip to content Information Technology at Purdue
Rosen Center for Advanced Computing

RCAC Resources

The Rosen Center for Advanced Computing (RCAC) maintains many different resources for computation and storage. Here is a brief introduction to our resources. More information and detailed documentation is available for each resource listed below.

• Coates

  The Coates community cluster was installed July 21, 2009. Coates is the largest academic entirely 10GigE cluster in the world. If you are interesting in buying into Coates, please fill out our online order form.

• Steele

  The Steele community cluster was installed in May, 2008, in an unprecendented single-day installation. It replaces and expands upon RCAC resources retired at the same time, including the Hamlet, Lear, and Macbeth clusters. Steele consists of 893 8-core Dell 1950 systems with various combinations of 16-32 GB RAM and Gigabit Ethernet and Infiniband. The large amount of memory in this system makes it well suited for large parallel jobs, providing fast communication between processors via shared memory. With 8 cores per node, sizable multithreaded programs (parallel codes that do not use MPI) should also be a good fit for Steele.

• Caesar

  Caesar is an SGI Altix 4700 system. This large memory SMP design features 128 processors and 512 GB of RAM connected via SGI's high-bandwidth, low-latency NUMAlink shared-memory interface. The extremely large amount of shared memory in this system makes it ideal for jobs where many processors must all share a large amount of in-memory data, and for large parallel jobs, using shared memory for fast communication between processors.

• Moffett

  Moffett is a SiCortex 5832 system. It consists of 28 modules, each containing 27 six-processor SMP nodes for a total of 4536 processors. The SiCortex design is highly unusual; it pairs relatively slow individual processors (633 MHz) with an extraordinarily fast custom interconnect fabric, and provides these in very large numbers. In addition, the SiCortex design uses very little power and thereby generates very little heat. Moffett is best suited to very wide parallel jobs with very high communication needs and may be used to explore the scalability of parallel algorithms. Serial applications, on the other hand, would suffer from the individually slow processor speeds and are not recommended.

• BoilerGrid (Condor Pool)

  BoilerGrid (Condor Pool) is a large, high-throughput, distributed computing system provided by RCAC and using the Condor system developed by the Condor Project at the University of Wisconsin. BoilerGrid (Condor Pool) provides a means for users to run programs on large numbers of otherwise idle computers in various locations, including both high-performance resources momentarily under-utilized and desktop lab machines not currently in use. Whenever a local user or scheduled job needs a given machine, the Condor job is stopped and sent to another Condor node as soon as possible. Because this model limits the ability to accomplish parallel processing and communications, RCAC decided to limit access to smaller, serial jobs. Condor jobs can be submitted from most of the RCAC systems (Gray, Pete, Prospero, Radon, Rossmann, Steele, Venice). You may also install Condor on your own desktop machine, and submit from that.

• Brutus (FPGA)

  Brutus (FPGA) is an experimental FPGA resource provided by the Northwest Indiana Compuational Grid (NWICG) through the Rosen Center for Advanced Computing (RCAC). Brutus (FPGA) currently consists of an SGI Altix 450 with two SGI RC100 blades with two FPGAs each, for a total of 4 FPGAs. Using Brutus (FPGA) effectively requires careful code development in either VHDL or Mitrion-C, but can result in significant performance increases. BLAST has been benchmarked on Brutus (FPGA) at 70x typical general-purpose CPU performance.

• Radon

  The Radon cluster is composed of desktop PCs recycled from instructional computing labs. Radon is currently entirely 64-bit Dell systems with Intel Pentium4 or Xeon processors of various speeds and with memory configurations between 2 and 4 GB of RAM. Nodes are connected with either 100 MB or Gigabit Ethernet. The machines reclaimed from instructional labs are older, slower, and lack high-speed interconnects, so high-communication or sizeable multithreaded programs are not a good fit. Still, there are a fair number of machines, and some codes may be able to take advantage of these effectively.

• Black

  The Black cluster is Purdue's portion of the Indiana Economic Development Corporation (IEDC) machine at Indiana University, the IU portion of which is known as "Big Red". Black consists of 256 IBM JS21 Blades, each a Dual-Processor 2.5 GHz Dual-Core PowerPC 970 MP with 8 GB of RAM and PCI-X Myrinet 2000 interconnects. The large amount of shared memory in this system provides very fast communication between processors via shared memory, making this system ideal for large parallel jobs.

• Pete

  The Pete cluster is composed of two parts, one owned by Earth and Atmospheric Sciences (EAS) and the other by the Network for Computational Nanotechnology (NCN). Pete consists of 166 HP Dual-Processor Dual-Core DL 40 systems with either 8 or 16 GB RAM and Gigabit Ethernet. The large amount of memory in this system makes it well suited for large parallel jobs. With 4 cores per node, sizeable multithreaded programs (parallel codes that do not use MPI) would not be a good fit for Pete.

• Prospero

  The Prospero community cluster consists of 19 Dell Quad-Processor 2.33 GHz Intel Xeon systems with 8 GB RAM and both Gigabit Ethernet and Infiniband interconnects. Each node has enough memory to run most jobs, and the high-speed Infiniband interconnect helps with many communication-bound parallel jobs. Compared to systems like Steele or Caesar, Prospero nodes have fewer processors per node, meaning that sizable multithreaded programs (parallel codes that do not use MPI) will be a poor fit.

• Venice

  Venice is a small cluster of Sun x4600 systems consisting of two front-end nodes and three compute nodes. The front-end nodes are both a Quad-Processor Dual-Core AMD Opteron 2216. The compute nodes are each an Eight-Processor Dual-Core AMD Opteron 8220 with 128 GB of RAM. The large amount of shared memory in this system makes it ideal for large parallel jobs, using shared memory for fast communication between processors. Compared to systems like Caesar, Venice cluster nodes have fewer processors per node, so sizable multithreaded programs (parallel codes that do not use MPI) will be a poorer fit.

• Gray

  The Gray cluster is solely a development platform to be used alongside the Indiana Economic Development Corporation (IEDC) machine Black. Gray is a place to compile code (mostly serial Condor applications) that is to be run on Black. Black is currently housed with Indiana University's "Big Red" system in Bloomington, Indiana. However, Gray is located on Purdue's West Lafayette campus. Gray includes a front-end server, several worker-node blades, and extra front-end hosts for campus and TeraGrid Condor users.

• Gold

  Gold is a small IBM Power5 system consisting of one front-end node and one compute node. The compute node is a Dual-Processor 1.5 GHz Dual-Core Power5 520 with 8 GB of RAM. This system is designed to be used only by users who have legacy IBM architecture-optimized or AIX-specific code. Gold is to help facilitate the porting of any code designed for the IBM SP system at Purdue, which was retired in 2008.

• Fortress (DXUL)

  The Fortress (DXUL) DXUL system is a large, long-term, multi-tiered file caching and storage system utilizing both online disk and robotic tape drives. Fortress (DXUL) was upgraded in the fall of 2006. It currently consists of an IBM p570 with four 1.65 GHz Power5 processors, 8 GB of RAM, and a 2.5 TB RAID disk cache with an effective capacity of 1.7 TB. Fortress (DXUL) also uses an ADIC Scalar 10K robotic tape library with a capacity of 1.2 PB (36 LTO-II drives and 2,000 LTO-II tape cartridges).

• Network Storage (BlueArc)

  Network Storage (BlueArc) (The BlueArcs) are currently two large, scalable, redudant hardware (FPGA) NFS servers (no operating system), providing extremely fast and reliable data storage and access. RCAC uses these to provide home directory, applications, scratch, and miscellaneous space to nearly all of our computational resource systems and servers. Network Storage (BlueArc) combined currently contain more than 240 TB of redundant, high-availability SATA and Fibre Channel disk space, and are a central component of RCAC's systems infrastructure.

  File-storage is two-tiered. New files and files in use are on fast fibre disks, while older files are migrated to slower SATA disks. This approach enables efficient utilization of storage resources while still providing both the performance and capacity required.