Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

SparkDSLR and GpuKMeans not working in cluster mode #71

Open
a-agrz opened this issue Nov 2, 2017 · 2 comments
Open

SparkDSLR and GpuKMeans not working in cluster mode #71

a-agrz opened this issue Nov 2, 2017 · 2 comments

Comments

@a-agrz
Copy link

a-agrz commented Nov 2, 2017

Hi !
I was testing the new merge of GPUDataset in GPUEnabler and got the following problems
I did my tests in a node that has 3 GPUS attached to it.

In Local mode, both examples SparkDSLR and GpuKMeans works fine

SparkDSLR in local mode:
image

GpuKMeans in local mode:
image

But I got this errors when executing SparkDSLR example in cluster mode,
image

and this errors when executing GpuKMeans example in cluster mode
image

PS: SparkGPULR (RDD version) works fine in sluster mode

So why datasets don't work in cluster mode? what I'm I missing here? is there some configurations that I need to do ??

Best regards

Aguerzaa

@josiahsams
Copy link
Member

The exception is due to the fact that, the kernel file(*.ptx) file, which is part of the examples.jar file, is not part of the driver/executor classpath and getClass.getResource returns NULL. So to make it part of the classpath, add the following entries in the Spark's conf/spark-default.conf file,

 spark.executor.extraClassPath	  /home/joe/GPUEnabler/gpu-enabler/target/gpu-enabler_2.11-1.0.0.jar:/home/joe/GPUEnabler/examples/target/gpu-enabler-examples_2.11-1.0.0.jar
 spark.driver.extraClassPath	  /home/joe/GPUEnabler/gpu-enabler/target/gpu-enabler_2.11-1.0.0.jar:/home/joe/GPUEnabler/examples/target/gpu-enabler-examples_2.11-1.0.0.jar
 spark.jars			  /home/joe/GPUEnabler/gpu-enabler/target/gpu-enabler_2.11-1.0.0.jar

@a-agrz
Copy link
Author

a-agrz commented Nov 3, 2017

Thank YOU @josiahsams; It's working now, here are the results I found:

SparkDSLR in cluster mode:
image

GpuKMeans in cluster mode:
image

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants