Options

how can I force rapidminer to use GPU when running usual operators?

mansourmansour Member Posts: 26 Contributor II
in Help
Dear All
I am running select attribute operators with a large dataset, but Rapidminer failed to finish the job and quit due to RAM problem.
I added a GPU  card to my computer and run it again, but Rapidminer didn't use GPU and show again the same error. 
Anywhy to force Rapiminer to use GPU?"
Regards.

Answers

  • Options
    tftemmetftemme Administrator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, RMResearcher, Member Posts: 164 RM Research
    Currently only the Deep Learning extension supports GPU executions.

    If you have problems with the memory and large datasets, try to split your data into chunks, perform the Select Attributes operator and append them later.

    It also can be that your process is set up in a way that you need an extreme amount of memory (for example if you perform the Select Attribute in a parallel executed Loop or Cross Validation. Sometimes it can help to look into the process and try to simplify the process to reduce the memory footprint while execution.

  • Options
    mansourmansour Member Posts: 26 Contributor II
    Dear @tftemme
    Thanks. What do you mean by splitting data into chunks? Raw or column splitting? 
  • Options
    tftemmetftemme Administrator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, RMResearcher, Member Posts: 164 RM Research
    I meant raw splitting. Maybe even reading in only chunks of the input data.
Sign In or Register to comment.