The Experimental Computing Laboratory (ExCL) was established in 2004 with the goal of providing application users and computer scientists with access to systems with leading edge architectures and software. ExCL is managed by the ORNL Future Technologies Group. ExCL researchers investigate architectures such as multi- and many-core processors, Graphics Processing Units (GPUs), non-volatile memory systems, and Field Programmable Gate Arrays (FPGAs). Most ExCL hardware is located in a large access-controlled server room in the JICS Building at ORNL (Building 5100), near the Future Technologies Group offices.
Open Compute Systems
- oswald.ftpn.ornl.gov: four nodes with each node containing
- Intel Xeon E52683 CPUs (x2)
- 256GB DDR4 Main memory
- Sandisk PX600 NVM appliance
- NVIDIA Tesla P100
- Nallatech 510T FPGA Card with (2) Altera Arria FPGAs (See these instructions for how to use the FPGAs)
- Mellanox FDR HCA
- Emu Chick system with novel migratory memory-side thread computing for applications with little data locality.
- megatron.ftpn.ornl.gov: two eight-core Intel Xeon E5-2643s
- newark.ftpn.ornl.gov: two NVIDIA GPUs (one Fermi C2050 and one Kepler K40c), two quad-core Intel Xeon E5520s
- atlanta.ftpn.ornl.gov: two AMD GPUs (one Tahiti Radeon HD 7990 and one Cypress FirePro v8800), two quad-core Intel Xeon E5520s
- peking.ftpn.ornl.gov: two quad-core AMD Shanghai processors
- madrid.ftpn.ornl.gov: two quad-core AMD Barcelona processors
- Kold: Four HP s250 nodes with two Intel Xeon E5-2670s and NVIDIA GPUs, connected with FDR Infiniband
If you have problems or want to make a request for software to be in this shared software repository, please send a message to firstname.lastname@example.org.