MSA

1
MSA2000Metacomputing Systems and Applications
2
Panel Session

• Metacomputing infrastructures
• Load balancing and scheduling
• Languages and environments
• Applications

3
Metacomputing infrastructures

• Clusters already here, widely available
• Need bigger pipes, OK
• Only challenge allocating and sharing
  resources administration -(
  

4
Load balancing and scheduling

• Static scheduling difficult to trade-off
  parallelism and communication even in the
  presence of unlimited resources
• Dynamic load-balancing master-slave techniques,
  use the past to predict the future''

5
Load balancing questions

• How to trade-off between data distribution
  parameters and process spawning or possible
  migration policies ?
• Redundant computations ?
• Anything new since PCs Ethernet ?

6
Languages and environments

• Who needs GLOBUS big machinery ?
• JAVA everything ?
• Transparency of resources ?
• Performance evaluation anything specific ?

7
Applications (1)

• All applications involving parallel
  computingPerformance problems due to
• Using a network of heterogeneous machines
• Relying on current (limited) programming
  environments
• Classical applications such as the grand
  challenges can be ported on metacomputing
  platforms
• Forget fine-grain parallelism deep
  hierarchy between all memory and communication
  layers
• Code coupling nicest application for
  metacomputing

8
Applications (2)

• Other applications out of the world of numerical
  (or scientific) computing
• data-bases, decision-support systems
• all kind of multimedia servers (PPI project at
  Caltech)
• Best candidates loosely-coupled applications
• All kinds of decomposition (functional, pipeline,
  data-parallel, macrotasking, ...)
• Actual challenge implementation of
  tightly-coupled applications

9
Applications questions

• NPACI or GrADS representative ?
• Who really needs the Grid ?
• LAPACK benchmark on the Grid ?