GraphX parameters tuning

[muaz-32 <mu...@euranova.eu>]

Hi everyone. 
I am doing my master thesis in the topic of Automatic parameter tuning of
graph processing frameworks. Now, we are aiming to optimize GraphX jobs. I
have an initial list of parameters which we would like to tune:
spark.memory.fraction
spark.executor.memory
spark.shuffle.compress
spark.default.parallelism
spark.shuffle.file.buffer
spark.io.compress.codec
spark.speculation
Plus some parameters related to java GC. 
I would like to hear your feedback: If anyone thinks that there's a missing
parameter which should be tuned or if there is a parameter in the list which
is not very important.

Best,
Muaz Twaty
EURA NOVA



--
Sent from: http://apache-spark-user-list.1001560.n3.nabble.com/

---------------------------------------------------------------------
To unsubscribe e-mail: user-unsubscribe@spark.apache.org