| Feature |
Benefit |
| Modular, functional blade design 2 options in compute blades, plus I/O and memory expansion graphics and RASC™ blades |
- Unlimited flexibility and fine-grained control over system configuration
- Improved performance density
- Optimize for any workload
|
| Compatible with future technology Socket compatible with future Intel Itanium processors, including dual core CPU's |
- Investment protection
- Expand on demand
- Easily upgrade or service individual components
|
| Scalable System Size Scales to 512 sockets or 1024 cores system size and as much as 128 TB globally addressable memory |
- Superior sustained performance for any programming model or workload
- Breakthrough potential for data-intensive applications
- Simplified programming
- Easy, low-cost administration
|
NUMAlink 4 interconnect, MPI support- 6.4 GB/s bandwidth, <1 microsecond latency
- SGI Message Passing Toolkit
|
- Industry-leading cluster application efficiency to hundreds of processors or more
- Scale out to hundreds or thousands of CPU cores while retaining global memory address
|
SGI RAS Roadmap- Fault prevention, detection, recovery
- Special focus on memory RAS
- Serviceability of blades
|
- Linux performance and choice without compromised reliability
- Maximum uptime, higher utilization
|
Standards-based design
Commodity CPU's, memory, I/O |
- Leading price-performance versus high-end RISC/Unix servers
- SGI-level quality at affordable cost
|
Certified on Industry-standard Linux®
Choice of SUSE or RedHat®SGI supplied and supported |
- Easier system bring-up
- One-stop support
|
