Silicon Graphics ICE X, User Manual

Page: 45 / 142

007-5806-004 25
Chapter 3
3. System Overview
This chapter provides an overview of the physical and architectural aspects of your SGI Integrated
Compute Environment (ICE) X series system. The major components of the SGI ICE X systems
are described and illustrated in the following sections:
• “System Models” on page 26
• “SGI ICE X System and Blade Architectures” on page 29
• “System Features and Major Components” on page 38
Because the system is modular, it combines the advantages of lower entry-level cost with global
scalability in processors, memory, InfiniBand connectivity and I/O. You can install and operate
the SGI ICE X series system in your lab or server room. Each 42U SGI rack holds one or two
21U-high (blade enclosure pairs). An enclosure pair is a sheetmetal assembly that consists of two
18-blade enclosures (upper and lower). The enclosures used in D-rack configurations are
separated by two “shelves.” In D-rack systems the shelves each hold three power supplies (shared
by the blade enclosures). In M-rack systems 1U shelves are used only to hold the blade enclosure
pairs and the power supplies are positioned on the sides of the blade enclosures. Each blade
enclosure also has an internal InfiniBand communication backplane. The 18 compute blades
supported in each enclosure can use one or two node boards, with ASICs, processors, memory
components and I/O chip sets mounted on them. The blades slide directly in and out of the
enclosures. Every compute node in a blade contains four or eight dual-inline memory module
(DIMM) memory units per processor socket. Optional hard disk or solid-state (SSD) drives and
MIC or GPU option boards are available. Each compute blade supports one or two individual node
boards. Note that a maximum system size of 72 compute blades per rack is supported at the time
this document was published. Optional chilled water cooling may be required for large
processor-count rack systems. Contact your SGI sales or service representative for the most
current information on these topics.
The SGI ICE X series systems can run parallel programs using a message passing tool like the
Message Passing Interface (MPI). The SGI ICE X blade system uses a distributed memory scheme
as opposed to a shared memory system like that used in the SGI UV series of high-performance
compute servers. Instead of passing pointers into a shared virtual address space, parallel processes
in an application pass messages and each process has its own dedicated processor and address
space.