High Performance Computing Systems
Please follow HPC Notices on Twitter - They are specifically designed to keep OSC clients up to date on system downtimes, outages, maintenance and software updates. You can can also see a full list of current system notices here.
Oakley, the Ohio Supercomputer Center HP Intel Xeon Cluster
 OSC’s new $4.1 million HP-built, Intel® Xeon® processor-based supercomputer, dubbed the Oakley Cluster, will feature more cores (8,328) on half as many nodes (694) as the center’s current flagship system, the IBM Opteron 1350 Glenn Cluster. The Oakley Cluster can achieve 88 teraflops, tech-speak for performing 88 trillion calculations per second, or, with acceleration from NVIDIA® Tesla™ graphic processing units (GPUs), a total peak performance of 154 teraflops.
Detailed system specifications:
- 8,328 total cores
- 12 cores/node & 48 gigabytes of memory/node
- Intel Xeon x5650 CPUs
- HP SL390 G7 Nodes
- 128 NVIDIA Tesla M2070 GPUs
- QDR IB Interconnect
- Low latency
- High throughput
- High quality-of-service.
- Theoretical system peak performance
- GPU acceleration
- Total peak performance
- Memory Increase
- Increases memory from 2.5 gigabytes per core to 4.0 gigabytes per core.
- Storage Expansion
- Adds 600 terabytes of DataDirect Networks Lustre storage for a total of nearly two petabytes of available disk storage.
- System Efficiency
- 1.5x the performance of current systems at just 60 percent of current power consumption.
Using Oakley, the HP Intel Xeon Cluster
Glenn, the Ohio Supercomputer Center IBM Cluster 1350
Phase One Glenn has been decommissioned December 14, 2011.
 The
Ohio Supercomputer Center's IBM Cluster 1350, named "Glenn", includes
AMD Opteron multi-core technologies and the new IBM cell processors.
The system offers a peak performance of more than 75 trillion floating
point operations per second and a variety of memory and processor
configurations. OSC's new supercomputer also includes blade systems
based on the Cell Broadband Engine processor. This will allow Ohio
researchers and industries to easily use this new hybrid HPC
The current hardware configuration consists of the following:
877 System x3455 compute nodes - Decommissioned December 14, 2011- 650 System x3455 compute nodes
- Dual socket, quad core 2.5 GHz Opterons
- 24 GB RAM
- 393 GB local disk space in /tmp
88 System x3755 compute nodes - Decommissioned December 14, 2011- 8 System x3755 compute nodes
- Quad socket, quad core 2.4 GHz Opterons
- 64 GB RAM
- 188 GB local disk space in /tmp
- Voltaire 10 Gbps PCI Express adapter
- 4 System x3755 login nodes
- Quad socket 2 dual core 2.6 GHz Opterons
- 8 GB RAM
- All connected together by 10 Gbps or 20 Gbps Infiniband
There are 36 GPU-capable nodes on Glenn, connected to 18 Quadro Plex S4's for a total of 72 CUDA-enabled graphics devices.Each node has access to two Quadro FX 5800-level graphics cards.
- Each Quadro Plex S4 has these specs:
- Each Quadro Plex S4 contains 4 Quadro FX 5800 GPU's
- 240 cores per GPU
- 4GB Memory per card
- The 36 compute nodes in Glenn contain:
- Dual socket, quad core 2.5 GHz Opterons
- 24 GB RAM
- 393 local disk space in '/tmp'
- 20Gb/s Infiniband ConnectX host channel adapater (HCA)
Using Glenn, the IBM 1350 Opteron Cluster at OSC
Ohio Supercomputer Center BALE Cluster
 The OSC BALE Cluster is a distributed/shared
memory hybrid system constructed from commodity PC components running
the Linux operating system.
The current hardware configuration consists of
the following:
The BALE Theater Cluster consists of 55 Workstation nodes and one login node configured with:
- Motherboard: ASUS M2NPV-VM
- graphics: nVIDIA GeForce 6150+nForce 430
- Ethernet: gigE on board (login node has an extra gigE card)
- Processors: AMD Athlon 64 X2 DualCore EE 4200+ (2.2GHz) AM2 Socket L2=512KB
- RAM: 4x1 GB PC5300 ECC DDR2 667MHz
- Infiniband: PCI-Express (8x) Single Port 4X Infiniband HCA Card - MemFree
- Hard Drive: 250GB SATA II Hard Drive 7200RPM w/ 16MB Cache
- login node: 2 x 500GB SATA II Hard Drive 7200RPM w/ 16MB Cache
The GPGPU/Visualization cluster consists of 18 nodes, each contains:
- two AMD 2.6GHz Dual-Core Opteron CPU's
- 8 GBs of RAM
- 750 GB SATA hard disk
- two NVIDIA Quadro FX 5600 Graphics cards
- Infiniband Dual Port HCA card
- Each Quadro FX 5600 card has 1.5GBs of onboard, high-speed memory and is capable of 330 peak Gflops
Using the BALE cluster at OSC
Ohio Supercomputer Center Mass Storage Environment
 The Mass Storage Environment (MSE) at OSC consists of servers, data storage
subsystems, and networks providing a number of storage services to OSC HPC
systems. The current configuration consists of:
- Cisco MDS9509 Storage Directors with Fibre Channel and iSCSI blades
- IBM FastT-900 storage servers
- Hitachi AMS1000 storage
- DataDirect Networks 9900 storage
- local disk storage on each compute node
- One IBM 3584 tape robot:
- Four Fibre Channel links.
- Eight LTO tape drives.
- 400 TB (raw capacity) of LTO tapes.
-
The servers for the home directories are 18 x3650s with the following:
- 16GB memory
- 2 Intel Xeon E5440 2.83GHz quad-core CPUs
- 2, 1 Gb/s Ethernet network interfaces
- 4Gb/s Fibre-channel host bus adapter
Home Directory Service
The MSE provides common home directories across all OSC HPC systems.
Each userid has a quota of 500GB storage and 1000000 files for its home
directory tree.
For projects that require more than 500GB storage and/or more than
1000000 files, additional storage space is available. Principal
Investigators should contact the OSC consultation service about
procedures for making additional storage available in the "project"
space outside the home directory.
Note: To transfer files to/from OSC systems, use scp or sftp between
your system and any of the HPC login nodes. Since home directories and
"project" space are shared across all systems, you can use any system
to transfer files.
The local storage on each compute node is used for scratch space
(/tmp) and spool space for stdout and stderr from batch jobs.
Some nodes have large scratch space. See the usage sections for
each cluster for additional information.
Parallel File System Service
The MSE provides a parallel file system for use as high-performance
shared scratch space using PVFS2. The current capacity of the PVFS2
subsystem is about 100TB.
Backup Service
All files on the home directories and on the "project" space are
backed up daily. Two copies of files in the home directories are
written to tape in the tape library. A single copy of files on the
"project" space is written to tape. Files on the local scratch (/tmp)
and on PVFS are not backed up.
|
