 |
|
 |
       |
 |
|
Mellanox Cluster Center offers an environment for developing, testing, benchmarking and optimizing products based on InfiniBand and 10 GigE technology. The center, located in Santa Clara, California, provides on-site technical support and enables secure sessions onsite or remotely.
The Cluster Center provides a unique ability to access the latest InfiniBand technology, even before it reaches the public availability. It provides a development testing and tuning environment for applications to take advantages of Mellanox’s superior characteristics.
Mellanox clusters utilize “Fat Tree” (or Constant Bisectional Bandwidth - CBB) network architecture to construct non-blocking switch configurations. Fat Tree network is a switch topology in which integrated non-blocking switch elements (crossbars) with a relatively low number of ports are used to build a non-blocking switch topology supporting a much larger number of endpoints. Using full Fat Tree networks is a key ingredient to deliver non-blocking bandwidth for high performance computing and other large scale compute clusters. 
Current Available Clusters
Neptune
- Mellanox ConnectX 20Gb/s InfiniBand technology
- 32 Colfax 1U 2P Dual Core AMD Opteron™ Cluster Nodes
- 64 AMD Dual-Core Opteron™ Model 275 processors
- Light weight 30 AWG InfiniBand cables from W. L. Gore & Associates, Inc.
- 6GB memory per node
- 425.9Gigaflops (Linpack Measured)
Helios
- Mellanox ConnectX 20Gb/s InfiniBand technology
- 32 Rackable Systems c1000 DC powered rack-mount servers
- 64 Quad-Core Intel® Xeon® 5300 Series Processors
- Light weight 30 AWG InfiniBand cables from W. L. Gore & Associates, Inc.
- 8GB FBD host memory from WinTec Industries
Vulcan
- Mellanox ConnectX 20Gb/s InfiniBand technology
- 32 Colfax 1U 2P Quad Core AMD Opteron™ Cluster Nodes
- 256 AMD CPU cores
- Zarlink's ZL60615 ZLynx active optical cable
- 4GB memory per node
 |
|