Hi Team,
I have created a model which is having large volume of SKUs. I have used the 3D and process flow approach both. Now, I am running a model using optimizer for 30 days and it is taking around 20 hours to complete.
When I am checking performance profiler, it is showing the time of different nodes in miliseconds which is very less. How to interpret this data against 20 hours of run time?
Is there any way to get split of 20 hours of run time to analyze it better?
Thank you!
Hi @Ankur A3, was Jordan Johnson's answer helpful? If so, please click the "Accept" button at the bottom of their answer. Or if you still have questions, add a comment and we'll continue the conversation.
If we haven't heard back from you within 3 business days we'll auto-accept an answer, but you can always comment back to reopen your question.
Profiling is something that we add manually to code paths that seem likely to have a strong performance impact. It is possible we have missed an important code path.
Profiling is only meant to work while the model is running in the interactive session. It won't profile data during an optimization or experiment.
If you reset your model, start profiling, and then fast-forward your model to the stop time (the same stop time your optimization uses), then stop profiling, how does that compare? How long does it take, and how much time is reported in the performance profiler? If those two times are very different, we can investigate which code paths we missed.
Talking about the performance of the optimizer is a lot more difficult. You say you ran the optimizer for 20 hours. How many CPUs were you using? How much of your RAM was being utilized? How many total replications of the model did it run?
Of course, if you can provide your model, we can support you more fully.
Hi @Jordan Johnson,
Thank you for your answer!
I am sharing some details here while optimizer was running:
At 09:00
At 11:00 (After 2 hours)
CPU: Intel(R) Core(TM) i7-9850H CPU @ 2.60GHz 100% Utilized
RAM: 18 GB Utilized out of 32 GB
1 replication per scenario
Total 10 Scenarios as given in screenshot
Any guesses based on this info?
Thank you!
I expect we need to see your model to be able to assess its performance. You can switch this post to private if you need to prior to any upload.
We need to separate whether the model is just slow in general, or if it is somehow slower during optimization. Try this:
If that is also a long process, then use the profiler to determine why the model takes so long. If the model runs quickly, but the optimizer takes a while, send the model so we can investigate further.
Actually, model is slow in general as well. I am trying to improve it using different option and those are working. I will update on this once completed.
Also, I would like to know that
Is there any way that we can ensure the current system (Processor/RAM) is getting fully utilized while model is running?
Or
Where to improve in terms of system configuration? RAM/Processor if we want to make it faster?
Thank you!
15 People are following this question.
FlexSim can help you understand and improve any system or process. Transform your existing data into accurate predictions.
FlexSim is a fully 3D simulation software environment. FlexSim can be used to simulate any process in any industry.
FlexSim®, FlexSim Healthcare™, Problem Solved.®, the FlexSim logo, the FlexSim X-mark, and the FlexSim Healthcare logo with stylized Caduceus mark are trademarks of FlexSim Software Products, Inc. All rights reserved.
