VTK/Multicore Vision

1. Accelerate algorithms that need access to the whole dataset

• Examples:
  • Streamlines
  • Graph algorithms
• What is needed:
  • Multi***threading abstraction
    • vtkMultiThreader
    • What about OpenMP, Intel TBB?
  • Thread safe data structures
• Pros:
  • Scalable
• Cons:
  • Filters need to change

2. Pipeline parallelism

• Analyze the pipeline
• Assign different filters to different threads
• Pros:
  • Automatic performance improvement for all filters
• Cons:
  • It is only useful for:
    • Streaming
    • Branching pipelines

3. Streaming

• All types of data:
  • Structured data
  • Unstructured data
  • Text (documents)
  • ...
• How do we break the data?

4. How do we bring (1), (2) and (3) together?

• Do algorithms need to provide more meta***data?
  • I can work with pieces
  • I need the whole thing
  • I scale up to N processors?

5. What about interactive applications

• Using separate threads to run the pipeline and to control user interface/rendering