Parallel and High Performance Distributed Computing
-
e.g. high performance RMI
|
"Large Scale Scientific Objects" -- Interfaces for Computing Systems and Applications ("Scientific JDBC")
-
Computer Services Framework needed for developing interoperable user interfaces, universal tools and/or systems like Globus/Legion
-
Wrappers for "legacy" numerical codes (as described in JSPICE presentation on Feb 28 workshop session)
-
Component Architecture -- longer term
|
Implementation Issues
-
Difficult to change VM even though some think might be essential for good memory management
-
Coping with memory hierarchy
-
high performance threads and synchronization
|
|
Define set of relevant Benchmarks
|
