|
Processor 1 |
Choice of Intel Xeon or AMD Opteron |
Intel® Xeon TM
Processors 5x00 Series, featuring the new 5500 processor |
AMD Opteron 2400 Series: Six Core - up to 2.60 GHz |
Intel® Xeon TM Processors 5500 Series
|
|
Compute node |
ProLiant BL280c, BL460c G6 or
BL465c |
ProLiant DL4x170h, DL160 G6, BL2x220c G6, BL280c G6 and BL460c G6 |
ProLiant DL165 G6, DL585 G6 and BL465c G6 |
ProLiant SL2x170z
|
|
Visualization node |
|
HP xw8600 workstations, DL160 |
HP xw 9400 workstations, DL165, DL385 G6 |
DL160 G6
|
| Utility/Control node |
BL280c. BL460, BL465c or DLxxx |
ProLiant DL380 G6 as well as compute nodes listed above |
ProLiant DL385 G6 as well as compute nodes listed above |
DL160 G6 or DL380 G6
|
|
Number of nodes |
Max of 8 blades plus one
DLxxx control node |
Minimum: 1 Utility node, and 4 compute or visualization nodes
Maximum: Total of 1024 nodes (Utility + compute + visualization nodes)
Cluster Platform BL Series max: 512 nodes |
Maximum of 1024 nodes, more by request
|
|
Memory per node |
BL2x220c: up to 32 GB per server (64 GB per Blade)
BL280c up to 96 GB (DDR3)
BL460c: up to 192 GB
BL465c: up to 64 GB |
DL4x170h: up to 128 GB
BL2x220c: up to 32 GB per server (64 GB per Blade)
BL280c up to 96 GB (DDR3)
BL460c: up to 192 GB
DL160/DL380: up to 144 GB (DDR3)
BL480c: up to 64 GB
|
DL165/
BL465c: up to 64 GB
BL685c: up to 256 GB DL385: up to 128 GB
DL585: up to 256 GB |
SL2x170z: up to 128 GB
|
|
Disk/Media bays |
BL2x220c: 1 SATA Drive per server/2 per
blade
BL46Xc: 2 drives
SB40c: 6 drives |
DL4x170h: up to 8 SAS or SATA LFF drives
DL160/BL480c: 4 drives (SAS or SATA)
DL380: 16 SFF drives
BL2x220c: 1 SATA
drive per server/2 per blade
BL280c: 2 SATA, SAS or SSD drives
BL460c: 2 SATA or SAS drives
BL480c: 4 SATA or SAS drives |
BL465c/BL685c: 2 drives
DL165: 4 drives (SATA or SAS)
DL385: 16 drives (SATA or SAS)
DL585: 8 drives (SAS or SATA SFF)
|
SL2x170z: 1 SATA NHP dirve
|
|
Cluster interconnect2 |
4x DDR or 4x QDR InfiniBand and Gigabit Ethernet. |
|
Management network |
Gigabit Ethernet with 10/100
console network |
10/100 network or Gigabit Ethernet; 10/100 console network to node-based
embedded management interfaces. ProCurve switches |
|
SAN Storage in cab |
Optional SB40c Storage Blade |
Optional MSA2000 and MSA20/MSA30 Direct Attached
Optional SB40c storage blade per Server Blade |
Optional MSA2000 and MSA20/MSA30 Direct Attached |
|
Operating System options |
Linux: Red Hat Enterprise Linux and SUSE SLES
Microsoft Windows Server 2008 |
Linux: Red Hat Enterprise Linux and SUSE SLES |
|
Operating environment options |
HP-MPI across all operating systems |
|
Cluster management software options |
Linux: HP Cluster Management Utility (CMU), Platform HPC Workgroup Manager2
Windows: HPC Server 2008 |
Linux: HP CMU Platform HPC Workgroup Manager2
Windows: HPC Server 2008 |
Linux: HP CMU
|
|
Packaging:
|
The Cluster Platform
Workgroup System is offered as a standalone pedestal on wheels or in 22U rack
with a single c3000 enclosure with up to 8 blades, and optional head node in
rack. |
- High density packaging designed for optimal footprint and lowest entry
price—available up to 128 compute nodes (delivered in 4 cabinets) and up to 512
blade nodes (1024 nodes with BL2x220c)
- Modular packaging designed for expandability, with racks assembled as
Compute Building Blocks for compute nodes and network switches, Utility Building
blocks for control nodes and other cluster administration tasks, optional racked
monitor and keyboard, and Interconnect Building Blocks for high performance
interconnect switches
|
72 compute nodes per Compute Building Block (in 42U rack), integrated with Utility Building Block
|
|
Services |
HP Enhanced Services available for platforms. HP offers on-site field
installation. HP software support available for operating systems and HP
software options. |
|
Warranty |
Standard HW Warranties; software warranty specific to product options |
Printable
version
