Re: WEKA Processing Requirements

Just to reiterate that this depends on which algorithms you want to apply.
A lot of the algorithm implementations in WEKA are single-threaded so a
processor that achieves high single-thread performance can be useful. On
the other hand, you may be running a lot of experiments with many
algorithms in parallel, in which case you want as many processing cores as
possible anyway, regardless of whether the individual algorithms are single
threaded.

I don't have any hard numbers, but I have been very impressed by how fast
WEKA runs on the newer ARM-based Macs (I have access to an M1-based Mac
Mini). Note that we have made a release of the netlibNative package for
ARM-based Macs as well (
https://weka.sourceforge.io/packageMetaData/netlibNativeOSXarm/index.html).

There is a new Mac Mini with the M2 Pro chip out, see

  https://www.macrumors.com/2023/01/19/mac-mini-m2-pro-geekbench-scores/

It supports up to 32 GB of RAM. The M2 Max would be even better though, as
it seems to support up to 96 GB of RAM.

Of course, it all depends on what kinds of jobs you need to run. We have
one server computer with two 64-core AMD processors that has one 1TB of
RAM, and you can obviously run WEKA on that sort of machine as well.

Cheers,
Eibe

On Mon, 23 Jan 2023 at 14:02, Peter Reutemann <[email protected]>
wrote:

> > I just had a few questions about the ideal hardware requirements for
> running WEKA. We use it quite often for processing large numbers of
> samples, 1000+ samples with ~400 variables. We are looking into building a
> custom PC to run WEKA as efficiently as possible.
> >
> > Are there any hardware requirements we should be mindful of? Such as,
> would WEKA benefit from a powerful GPU? Any advice would be fantastic.
>
> It depends very much on your algorithms:
> - using locally weighted models? -> plenty of RAM
> - using kernel-based models? -> plenty of RAM
> - large ensembles? -> plenty of RAM
> - does the algorithm support multi-threading? -> add plenty of cores
> - some algorithms can take advantage of a GPU as well:
>   https://waikato.github.io/weka-wiki/speeding_up_weka/#gpu-acceleration
>
> If you are planning on adding more models, be generous in number of
> cores/RAM beforehand. :-)
>
> You can monitor your algorithm's memory consumption from the GUIChooser
> with:
> Program -> Memory usage
>
> A percentage is being displayed as tool tip when hovering over the
> graph. You can turn that percentage into a rough estimate via the
> "memory.max" property in the "SystemInfo" dialog (from the Help menu
> in the GUIChooser).
>
> Cheers, Peter
> --
> Peter Reutemann
> Dept. of Computer Science
> University of Waikato, Hamilton, NZ
> Mobile +64 22 190 2375
> https://www.cs.waikato.ac.nz/~fracpete/
> http://www.data-mining.co.nz/
> _______________________________________________
> Wekalist mailing list -- [email protected]
> Send posts to [email protected]
> To unsubscribe send an email to [email protected]
> To subscribe, unsubscribe, etc., visit
> https://list.waikato.ac.nz/postorius/lists/wekalist.list.waikato.ac.nz
> List etiquette:
> http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html
>

_______________________________________________
Wekalist mailing list -- [email protected]
Send posts to [email protected]
To unsubscribe send an email to [email protected]
To subscribe, unsubscribe, etc., visit https://list.waikato.ac.nz/postorius/lists/wekalist.list.waikato.ac.nz
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html