RockMountPro Clusters (Number of Scalable Compute Nodes per Standard 42U Rack Cabinet)
1 Rack
2 Rack
3 Rack
4 Rack
5 Rack
Number of Nodes
80
160
320
640
960
Number of CPUs
160
320
640
1,280
1,920
Number of CPU Cores
960
1,920
3840
7,680
11,520
Memory Capacity
7.68TB
15.36TB
30.72TB
61.44TB
92.16TB
System Performance
11.2TF
22.4TF
44.8TF
89.6TF
134.4TF
High-availability clusters (also known as Failover Clusters) are implemented primarily for the purpose of improving the availability of services that the cluster provides. They operate by having redundant nodes, which are then used to provide service when system components fail. The most common size for an HA cluster is two nodes, which is the minimum requirement to provide redundancy. HA cluster implementations attempt to use redundancy of cluster components to eliminate single points of failure.
Computational Density and Flexible Interconnect Options
The RockMountPro Clusters based on the RockMountPro system provides an extremely flexible and manageable way to create high-density computing. It is designed to scale, with little effort or re-configuration. For more flexibility, the RockMountPro Cluster system offers two subrack configuration options: the 5000 series offering up to 10 blade servers per subrack in a 5U form factor and the 8000 series offering up to 16 blade servers per subrack in an 8U form factor.
Configurations are now complemented with storage blade expansion options delivering up to 2TB or additional storage per node or GPU expansion options offering customers affordable compute and storage blade options in a single solution building block.
In addition, the RockMountPro Clusters offers a variety of interconnect options to include only standard Ethernet or Infiniband (IB) switches. RockMountPro offers flexibility for customers who may choose to deploy either an Ethernet or IB-based HPC Cluster with options for DDR or QDR configurations.
Cluster Management Options
RockMountPro offers configuration options to include RockMountPro Clusters Management Software or Rocks+ and MOAB. All options can be easily tested and pre-integrated as part of a complete package to include HPC professional services and support.
Software Rolls Package Choices
Intel Cluster Ready Roll
Intel Developer Roll (Compilers)
PGI Roll (Portland Group Compilers)
MOAB Roll (Platform)
LSF Roll (Platform)
TotalView Roll (Debugger)
CUDA Roll (Compliers)
Support Roll
Cluster Management Options
Workland Management, Cluster Management, and Access Portal benefits:
Platfor
management and monitoring
Remotely manage up to tousands of compute nodes through support for management and sub-management servers.
Security via user ID/Password using SSH/SHA
Web interface option
support for GigE and InfiniBand Interconnect configurations
RockMountPro Clusters Configuration
RockMountPro Blade Clusters
Blades
80 Blade nodes per RockMountPro Blade Clusters
Processors
160 based on intel Oeon Processor 5600 Series
Cores
960 processing cores per rack cabinet
System Performan
Up to 11.2TF
Memory Capacity
7.68TB mas. per system
Storage Capacity
Up to 160 internal 2.5" HDDs, equal to 80TB of local storage
Networking
Gigabit or InfiniBand high speed interconnects, networking, storage, operating system, cluster management software
Rack Configuration
Standard 42U/19" rack, 2U rack space available for switches
RAS features
Reliability, Availability, and Serviceability with shared power and cooling system design
Power Consumption
16kW to 32kW, depending on system configuration
Subrack System
Subrack System - 5000Series
Subrack System - 8000Series
Device Bays
Up to 10 blade nodes
Storage Expansion
Form Factor
5U rackmountable
8U rackmountable
CPU Capacity
Up to 20 CPUs
Up to 32 CPUs
CPU/GPU Option
Up to 5 CPU blades/ 5 GPU expansion blades
N/A
Power Supply
Up to four 1625W high-efficiency PSUs in redundant N+1 configuration Maximum power delivery - 4.707kW
Up to four 1625W high-efficiency PSUs in redundant N+1 configuration Maximum power delivery - 7.0605kW
Cooling
Up to 3 redundant cooling fan unit (CFU)
Up to 6 redundant cooling fan unit (CFU)
Ethernet I/O
Optional: Up to 2 GbE LAN ports for iSCB management communications
Optional: Up to 2 GbE LAN ports for iSCB management communications
Dimensions
8.75"H x 19"W x 26"D
14"H x 19"W x 26"D
Weight
173.6 lbs (78.4kg) max
304.9 lbs (138.6kg) max
Management
Support for high-speed interconnects, Rackmountpro Cluster Engine Management SW, Windows or Linux OS
Support for high-speed interconnects, Rackmountpro Cluster Engine Management SW, Windows or Linux OS
Compute Blad Node
RockMountPro Clusters
Processor
intel Xeon Processor 5600 Series
Processor Capacity
Two; Choice of 4 or 6 cores per processor
Chipset
Intel 5520 chipset
System Bus
Intel QuickPath Interconnect (QP)
Memory Type
800/ 1066/ 1333MT/S ECC RDIMM DDR3 memory
Memory Capacity
Up to 96GB in 12 DIMM sockets
Disk Controller
Intel I/O controller hub (ICH10R)
Drive Bays
Two fixed 2.5" SATA HDDs
Storage Capacity
1.0TB per blade
Storage Expansion
Storage expansion accessory with up to 4.0TB per expansion blade (optional)
Two USB 2.0 compliant ports, two RJ-45 LAN ports, VGA, serial, optional QSFP (IB)
Expansion Slots
One low-profile PCIe 2.0 x16 slot with riser card
Dimensions
5"H x 1.75W x 25"D (127mm x 44.5mm x 635mm)
Weight
11.5 lbs (5.2 Kg)
Temperature
Operating: 10 - 35° C
Remote Server Mgmt
IPMI 2.0 integrated on the BMC
Storage Expansion Blade Node
Storage Expansion Blade - ex100Pro
Storage Expansion Blade - ex200Pro
Configuration
Storage Expansion blade works side by side with the RockMountPro computer server.
HDDs are hot-swappable and connected directly to the compute blad via 4-port Raid Card.
Storage Expansion blade works side by side with the RockMountPro computer server.
HDDs are hot-swappable and connected directly to the compute blad via 8-port Raid Card.
Storage Capacity
Up to 2.0TB of storage capacity SATA/SAS HDDs per compute blade
Up to 4.0TB of storage capacity SATA/SAS HDDs per compute blade
1U Server Special
About Us
