Appeal No. 2002-0792
Application 08/801,646
independent form including all of the limitations of the base
claim and any intervening claims.
Invention
The invention relates to a method of scheduling parallel
processes in a distributed, multi-kernel, multiprocessor system.
In particular, the present invention relates to a system and
method for scheduling parallel processes with no kernel-to-kernel
communication. A thread model of program execution has proven to
be a viable method for parallel execution of program code both in
single and multiprocessor machines. Under the thread model,
programs are partitioned into a set of parallel activities. Each
activity during execution of the program code is called a thread.
See page 1 of Appellants’ specification.
The problem with the prior art systems is that if
scheduling of such a multi-threaded program requires excessive
kernel-to-kernel communication, the performance of the program
will suffer. Appellants’ invention solves this problem by using
nanothreads. A shared arena is provided in the user memory,
wherein the shared arena includes a register save area for each
of the plurality of threads. See page 9 of Appellants’
specification.
2
Page: Previous 1 2 3 4 5 6 7 8 9 10 11 12 13 Next
Last modified: November 3, 2007