FAQ for MaTEx Other ML Algorithms

abhinavvishnu edited this page Feb 20, 2017 · 1 revision

This FAQ is for all MLDM algorithms other than Deep Learning

Q: I want to use MaTEx on desktop/laptop. How should I select the number of processes to use?

A: Please run the following command, which works on most linux and mac based systems:

cat /proc/cpuinfo | grep processor | wc -l 

The output of this command should be the maximum number of processes that you should use.

Q. I want to use MaTEx on Amazon EC2. How should I get going?

A. Please look at the following documentation: http://mpitutorial.com/launching-an-amazon-ec2-mpi-cluster/ After successfully running simple MPI programs, please use the INSTALL instructions.

Q. When I run MaTEx, it asks for my password. How should I solve it?

A. To run MPI programs, it is always recommended to setup ssh keys. There are several tutorials to set these up, such as https://www.digitalocean.com/community/articles/how-to-set-up-ssh-keys--2

Q. When I run MaTEx, it hangs. How should I solve it?

A. MaTEx may hang, if the number of samples in the dataset is lesser than number of processes. For performance reaons, the number of samples should be significantly larger than number of processes. Please make sure that you can run a simple program such as:

mpirun -np 2 `echo hostname`. 

The actual number of processes can be changed to the number of processes you are using in the application.

Q. Does MaTEx use OpenMP?

A. At the moment, OpenMP is not used by any MaTEx program. This is likely to change in the future.

Q. Does MaTEx work on GPUs?

A. No. Currently, we have no plans to support GPUs.