Why is there a maximum number of workers limitation for parallel computing toolbox
17 views (last 30 days)
Show older comments
I love MATLAB software, I've been using it for a long time. It saves me tons of time for ease of documentation, debugging, and many built in features. Recently, however, I've started using the parallel computing toolbox and immediately found an ~ 8x speedup when moving from a dual core year 2004 P4 (non parallelized code) to a dual dual-core (4 cores total) year 2006 Xeon (parallelized code). This has led me to the decision to upgrade my workstation in the next 2-3 months to take even more advantage of this great speedup factor.
Unfortunately, I've found that MATLAB sets a maximum limit of 8 workers (basically, 8 cores) for utilization. It seems that you can get around this by getting the distributed computing toolbox, but that doesn't seem to be as available (I am a University license user). Furthermore, I resent the idea that I need to upgrade to a much more expensive toolbox for a functionality that seems like ought to already be there.
It seems to me that this is a gross limitation that MATLAB really needs to remove ASAP. This is the first time I've found a situation where MATLAB just plain can't accommodate the mainstream hardware out there right now. It's a bit ludicrous that you can now easily put together a 48 core workstation, and MATLAB would only be able to use 16% of your resources!
This limit needs to be lifted, and MATLAB needs to set this at the top of their priority list for updates and upgrades. But in case there might be something I'm missing, I'll go ahead and consider: is there any good reason why MATLAB has set this limitation? Or does everyone agree that this should be top of the priority list?
Cheers, MNR
0 Comments
Answers (1)
See Also
Categories
Find more on Cluster Configuration in Help Center and File Exchange
Products
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!