The current limit of workers in MATLAB2013b for a local cluster is 12 which doesn't reach the server limit of 32 logical cores. A MDCS license for 32 workers costs more than the server hardware. Will MathWorks correct these disparities in the future?
Show older comments
We get a cluster of 12 workers on our local machine for the cost of the Parallel Computing Toolbox. But we run MATLAB on servers with 32 logical cores. So we looked at the MATLAB Distributed Computing Server for 32 workers, which costs (much) more than the server hardware.
When will MathWorks correct this disparity?
Solutions might be releasing a new parallel pool for larger servers (16-64 local workers). Or dropping the price of MDCS to match the trend for more cores but slower cores. Or just increase the number of local workers released with PCT to 16 or 32?
Accepted Answer
More Answers (2)
As an end user, I can agree it's always nice to have lower prices. However, I don't know if the ratio of software to hardware cost is going to be a persuasive argument. Why shouldn't the software cost significantly more than the hardware? Why value the software less? The Parallel Computing Toolbox already costs significantly more than the GPU cards needed to run gpuArray...
Walter Roberson
on 29 Nov 2013
0 votes
Mathworks does not discuss product plans in public, especially not competitive pricing matters. You will need to talk to a sales representative, who might be able to get authorization to fill you in.
If, hypothetically, Mathworks had told me the plans, then I would be under NDA (Non Disclosure Agreement) to not reveal anything. "Those who can say, don't know. Those who know, can't say."
Categories
Find more on MATLAB Parallel Server in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!