topic Re: Question regarding hardware requirements in Discussions

Question regarding hardware requirements

PeggyTsai — Tue, 15 Jul 2025 10:33:21 GMT

Hello,

Recently we’re doing some heavy loading calculation. The time required is driving us crazy.

Because it’s both not fast enough to cope with our daily processing requirement.

For this issue, we had two sets of hardware ran a test on one random project for [Process Capability](116 Columns, 3968 rows):

With INTEL I7-14700 + 64GB Ram, costs 2.5 hrs
With AMD Ryzen 7975WX + 128GB Ram, costs 1 hr

So the questions would be:

From the above, running JMP(ver18, non Pro) with Windows system feels like depending on single core capacity. Is it?
Should the answer be Yes, what would be the recommended direction hardware upgrades?
Is JMP Pro running with multi-core, or is it the same with non-Pro versions? Would there be trial license available?
Is there difference between Intel or AMD processor?

Re: Question regarding hardware requirements

Craige_Hales — Tue, 15 Jul 2025 13:02:49 GMT

You should be able to use the Windows task manager to answer some of these questions. You should be able to tell if there's not enough memory and it's paging, and you should be able to tell if it's using more than one CPU.

Re: Question regarding hardware requirements

Paul_Nelson — Tue, 15 Jul 2025 13:03:46 GMT

By far, the amount of RAM is the most significant factor. Throw as much RAM into the system as you can. Remember JMP loads all the data in memory so you need to maximize the RAM if you are handling large data. Second, SSD drives wherever the system's swap space is located. When physical RAM runs out the system has to turn to swap space to handle the memory pressure. SSD are dramatically faster than spinning disks.

There are parts of JMP that are multithreaded, but that depends on the particular statistical platform.

Re: Question regarding hardware requirements

Jed_Campbell — Wed, 16 Jul 2025 16:39:34 GMT

In addition to what's already been mentioned by @Craige_Hales and @Paul_Nelson, can you clarify what you mean by "heavy loading calculation"? Are you running a script, or a built-in JMP Platform? Scripts can often be sped up considerably by approaching the problem differently. For example, I know a user that improved a process from 2 hours to just 15 seconds by using matrix calculations rather than table manipulation and using a built-in JSL function instead of doing a calculation via scripting.