Generic Parallel Machine Architecture

1
Generic Parallel Machine Architecture
Storage Hierarchy
Proc
Proc
Proc
Cache
Cache
Cache
L2 Cache
L2 Cache
L2 Cache
L3 Cache
L3 Cache
L3 Cache
potential interconnects
Memory
Memory
Memory

• Key architecture question Where is the
  interconnect, and how fast?
• Key algorithm question Where is the data?

2

• Columbia (10240-processor SGI Altix, 50
  Teraflops, NASA Ames Research Center)

3

• Beowulf (18-processor cluster, lab machine)

4
Cray XMT (highly multithreaded shared memory)
5

• AMD Opteron quad-core die

6
Multicore SMP Systems
7

• Michael Perrone (IBM) Proper Care and Feeding
  of Multicore Beasts
• http//www.csm.ornl.gov/workshops/HPA/documents/1
  -arch/feeding_the_beast_perrone.pdf

8
The nVidia G80 GPU

• 128 streaming floating point processors _at_1.5Ghz
• 1.5 Gb Shared RAM with 86Gb/s bandwidth
• 500 Gflop on one chip (single precision)

9
More Detail on GPU Architecture
10

• Top 500 List
• http//www.top500.org/list/2007/11/100