OGO 2.1 SGI Origin 2000

1
OGO 2.1SGI Origin 2000

Robert van Liere CWI, Amsterdam TU/e, Eindhoven
11 September 2001
2
unite.sara.nl

• SGI Origin 2000
• Located at SARA in Amsterdam
• Hardware configuration
• 128 MIPS R10000 CPUs _at_ 250 Mhz
• 64 Gbyte main memory
• 1 Tbyte disk storage
• 11 ethernet _at_ 100 Mbits
• 1 ethernet _at_ 1 Gbit

3
Contents

• Architecture
• Overview
• Module interconnect
• Memory hierarchies
• Programming
• Parallel models
• Data placement
• Pros and cons

4
Overview - Features

• 64 bit RISC microprocessors
• Large main memory
• Scalable in CPU, memory and I/O
• Shared memory programming model

5
Overview - Applications

• Worldwide /- 30.000 systems
• 50 with gt128 CPUs
• 100 with 64-128 CPUs
• 500 with 32-64 CPUs
• Computing serving many CPUs and memory
• Database serving many disks
• Web serving many I/O

6
System architecture 1 CPU

• CPU cache
• One system bus
• Memory
• I/O (network disk)
• Cached data

7
System architecture N CPU

• Symmetric multi-processing (SMP)
• Multi-CPU caches
• One shared bus
• Memory
• I/O

8
N CPU cache coherency

• Problem
• Inconsistent cached data
• Solution
• Snooping
• Broadcasting
• Not scalable

9
Architecture Origin 2000

• Node board
• 2 CPU cache
• Memory
• Directory
• HUB
• I/O

10
Origin 2000 Interconnect

• Node boards
• Routers
• Six ports

11
Interconnect Topology
12
Sample Topologies
13
128 Topology
14
Virtual Memory

• One CPU, multi programs
• Page
• Paging disk
• Page replacement

15
O2000 Virtual Memory

• Multi CPU, Multi progs
• Non-Uniform Memory Access
• Efficient programs
• Minimize data movement
• Data close to CPU

16
Latencies and Bandwidth
17
Application performance

• Scientific computing
• LU, ocean, barnes, radiosity
• Linear speedup
• More CPUs -gt performance

18
Programming support

• IRIX operating system
• Parallel programming
• C source level with compiler pragmas
• Posix Threads
• UNIX processes
• Data placement
• dplace , dlock, dperf
• Profiling
• timex, ssrun

19
Parallel Programs

• Functional Decomposition
• Decompose the problem into different tasks
• Domain Decomposition
• Partition the problems data structure
• Consider
• Mapping tasks/parts onto CPUs
• Coordinate work and communication of CPUs

20
Task Decomposition

• Decompose problem
• Determine dependencies

21
Task Decomposition

• Map tasks on threads
• Compare
• Sequential case
• Parallel case

22
Efficient programs

• Use many CPUs
• Measure speedups
• Avoid
• Excessive data dependencies
• Excessive cache misses
• Excessive inter-node communication

23
Pros vs Cons

• Multi-processor (128 )
• Large memory (64 Gbyte)
• Shared memory programming

• Slow integer CPU
• Performance penalty
• Data dependencies
• Off board memory