Parallel Processing is better than buying another MATLAB licence?

13 次查看(过去 30 天)
Hello,
I'm considering whether it's more beneficial to purchase 2 or more MATLAB licenses or to invest in the Parallel Processing Toolbox. How many GPUs can we utilize with the Parallel Processing Toolbox, and is it worth the investment? I would greatly appreciate insights from someone with experience in this matter. Thank you!

采纳的回答

Andreas Goser
Andreas Goser 2024-4-19
While I am not an expert myself, I have listened to a few similar discussions and the general statement was it depends on your application.
If you have the hardware, I suppose you can test your application with MATLAB / PCT trial licenses?
  3 个评论
Denizhan AKINCI
Denizhan AKINCI 2024-4-19
Indeed, I do have access to MATLAB with the Parallel Computing Toolbox (PCT) through my student account. However, since I'm currently interning at a company, they don't have a PCT license under their corporate account. That's why I inquired about the viability of purchasing additional licenses or just buying a new matlab license. I'll give the trial licenses a try and see what options are available. Thank you for your assistance!

请先登录,再进行评论。

更多回答(1 个)

Joss Knight
Joss Knight 2024-4-21
There may be a misunderstanding here. Your typical license is to a user or a computer. That user, or that computer, can open as many copies of MATLAB as you like. The advantage of Parallel Computing Toolbox is that you can coordinate all those MATLABs through a single client, such as running a hyperparameter optimization. MATLAB Parallel Server lets you do the same but on multiple computers using a single license - maybe that's what you're thinking of?
  2 个评论
Denizhan AKINCI
Denizhan AKINCI 2024-4-22
So my original question is this actually : Can I optimize the execution time of multiple concurrent simulations in MATLAB? I've noticed that running a second MATLAB instance while the first one is running significantly slows down both, extending the overall execution time. Would acquiring another license alleviate this issue, or should I explore the Parallel Computing Toolbox (PCT) as a solution ?
Joss Knight
Joss Knight 2024-4-22
Extremely unlikely, if your plan is just to run the same code but on each worker. Parallel workers prevent resource contention by being restricted (by default) to a single physical core, and then the hope is that all the MATLABs together will use all your cores. It sounds like MATLAB is already doing a good job of multithreading and using all your resources and that's why there's contention between each MATLAB.

请先登录,再进行评论。

类别

Help CenterFile Exchange 中查找有关 Startup and Shutdown 的更多信息

产品


版本

R2023b

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by