Next: Efficiency of parallel computation
Up: Parallel execution
Previous: Experimentation of the maximum
Figure 1.6:
The speed up effect of DeepForest on shared memory
system (Cray CS-6400).
 |
The speed up effect in parallel computation is shown in Figure
1.6. We also carried out estimation of branch lengths in
10 MRF family genes (Figure 1.7). In many cases, branch
lengths are greater in the maximum likelihood method than in the
neighbor joining method. Especially, in the case of internal nodes, the
differences are obvious.
Figure 1.7:
Comparison of branch lengths (substitution/site) between the
neighbor joining method (above numbers) and the maximum likelihood
method (below numbers).
 |
Satoshi OOta
1999-03-06