contents up previous next
Up: Results
Previous: Single Processor Timings
Next: Analysis

Multiprocessor Timings

 figure88
Table 2: Timings for multiprocessor isosurface extraction of three data sets, using Marching Cubes, edge-hashing and cache-ring algorithms, and varying numbers of CPU's.

Table 2 contains timing results from running the three isosurfacing methods on the three larger data sets and using varying numbers of processors. Once again we have included separate timing values indicating the amount of overhead in the hash table and cache-ring runs. All timings have been summed over the utilized processors. Additionally, we have included a separate ``Wall Clock'' timing, to demonstrate the amount of real-world time that passed during execution.



Scientific Computing and Imaging
Wed Apr 23 17:39:06 MDT 1997