A Macrotask-level Unlimited Speculative Execution on Multiprocessors (16/19)

The following description is based on the paper presented at 1995 ACM International Conference on Supercomputing head in Spain on July 3rd to 7th,1995.

If you want to have the copy of the paper, please let me know.


back index next


Now, I'd like to address myself to the most important part. "Evaluation of the distributed control scheme". The speedup in comparison with the normal execution - I mean without speculation - is presented.
The target machine is the EM-4 Multiprocessor consisting of 80 PEs whose clock cycle is 12.5MHz. In our implementation, the point to point communication costs 3.3us. and Broadcast costs 26.4us.
This figure illustrates the implementation of the distributed control on the EM-4. Four kinds of threads are programmed and running at every processor.