KNIME Analytics Platform - CPU 100% in Windows

KNIME Analytics Platform
1

Hi,

In terms of speed of execution, what’s best system configuration recommended for running KNIME in Windows.

Based on the documentation, understand that KNIME analytics platform can be installed in following OS:

Windows

  • Windows 10, 11

  • Windows Server 2016, 2019, 2022

However when KNIME is opened in Windows Server 2022, CPU % utilization reaches 100% even when no workflow is executed.

Please share some workaround to reduce this CPU utilization %

Thank you.

2

Hey @balabn23,

the reason for this is likely that the machine has no access to any GPU. We recently switched to a more performant rendering technique that seems to bring CPU to its limits if there is no GPU existing at all. You can switch to the old rendering mode in the preferences. (select: “SVG rendering (legacy)”).

What is the use case you are running the Analytics platform on a windows server?

image
image1718×1170 86.9 KB

Greetings,

Daniel

3 Likes
3

Thank you, Daniel. What is limitation of not WebGL rendering and shifting to SVG rendering. Will it impact performance where we have 64 GB RAM with 16 core logical processor. Also, can KNIME analytics software connect to Databricks / Azure database (or) do I have to move to KNIME pro / team / business hub mandatorily?

4

Hey @balabn23,

you might run into slower application behaviour in case of very large workflows. Can’t tell exact numbers, but with your specs I wouldn’t worry too much.
All nodes are freely usable without a KNIME Pro / team / business hub setup/account, but our commercial offerings will help you when you try to “productionise” your workflows. Lets say you want it to run every Monday or similar.

Greetings,

Daniel

1 Like
5

Thank you, Daniel. I would like to know what is best configuration (how many cores, model and RAM etc.) for KNIME to run at better speed. I am trying to install in Windows Server as a use-case where I can use batch executor commands to schedule the workflows.. I would also like to know what is the advantage of moving to KNIME pro / team / business hub over installation and running workflows locally in Windows Server 2022. Please share your views on best configuration for KNIME and disadvantage using Windows Server locally over KNIME Pro / team / business hub.

6

Hi Bala, I am Sandeep- Account Manager at KNIME. I’d be able to answer some of the questions for you. Can you please shoot an email to sandeep.ravi@knime.com, so that we could set up a time to connect?
Thanks and regards.

1 Like
7

keep in mind to share the information here or append the documentation about those recommendations to the questions raised by @balabn23 as this will benefit the users of the AP

2 Likes
8

@sandeepravi any news regarding the knowledge sharing?

9

Hey @fe145f9fb2a1f6b,

Let me jump in — I think this deserves a bit more detail.

At a high level the difference between the product tires comes down to this:

  • Batch executors are best suited for internal automation, if you have the skills and time for infrastructure management relying on community support only.
  • Pro / Team is the built in native way for scheduling and monitoring workflow runs, with some added benefits of extra features (shared insights, collaboration, credential management, etc). Real difference between them is Pro is designed for a single user while Team excels at collaboration.
  • Business Hub is designed to fulfill enterprise level needs, like, governance, centralized user management or setting up your own AI backends.

You can find a more detailed comparison here: KNIME Pricing | KNIME

Performance configuration

  • This depends heavily on your specific use case. There are minimum requirements, but beyond that it varies with workload and expectations.
  • Regarding the original issue in this thread (slow rendering in VPN setups): good news — the framework we use for workflow rendering recently released an update aimed at improving performance when no GPU is available. We’re currently monitoring how this matures, and if it holds up, it should significantly improve performance in those scenarios as well.

If you share a bit more about what you’re trying to achieve, I’m happy to point you in the right direction — and if it helps, I can also connect you with someone from our sales team to talk through your use case.

Greetings,
Gergely

4 Likes
10

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.