User 3069 | 1/24/2016, 11:41:26 AM
First I do:
import graphlab as gl gl.setruntimeconfig('GRAPHLABDEFAULTNUMPYLAMBDAWORKERS', 32) gl.setruntimeconfig('GRAPHLABDEFAULTNUMGRAPHLAMBDA_WORKERS', 32)
after that I start data processing and soon get this warning:
PROGRESS: Using default 16 lambda workers. PROGRESS: To maximize the degree of parallelism, add the following code to the beginning of the program: PROGRESS: "graphlab.setruntimeconfig('GRAPHLABDEFAULTNUMPYLAMBDAWORKERS', 64)" PROGRESS: Note that increasing the degree of parallelism also increases the memory footprint.
Indeed, by CPU usage, it's only 16 cores used.
I'm not sure, but this looks as a bug. GraphLab was updated less then a week ago.