Error with parpool (Error using parpool (line 103) Not enough input arguments.)

Hello,

since today I've got a problem with starting the parallel pool. I've used it very often in the past to train neural networks (the last time was yesterday) and it worked without any problems. Today I wanted to test some new parameters with my neural network and tried to train it but everytime when I try to start the parallel pool (for example from the command line) there is the error:

>> parpool
Starting parallel pool (parpool) using the 'local' profile ... Error using parpool (line 103)
Not enough input arguments.

Error using parpool (line 94)
Not enough input arguments.

I've nothing changed in my 'local' profile. The same error happens when I try to 'Validate' in the 'Cluster Profile Manager'. Here are the validation results:

VALIDATION DETAILS
Profile: local
Scheduler Type: Local

Stage: Cluster connection test (parcluster)
Status: Failed
Description:The validation stage encountered a MATLAB exception.
Command Line Output:(none)
Error Report:
Not enough input arguments.
Debug Log:(none)

Can someone help me?

Thanks.

[Edric Ellis]

Did you change something on your MATLAB path? I suggest trying to restore a default MATLAB path to see if that fixes things.

[Christoph Schöne]

I've just tried 'restoredefaultpath' but the error still occurs. I've found another post http://www.mathworks.com/matlabcentral/answers/196549-failed-to-start-a-parallel-pool-in-matlab2015a with the same problem. They suggest to use the command 'distcomp.feature( 'LocalUseMpiexec', false )' but this doesn't solve my problem, too. It's still the line 103, not enough input arguments error.

[Edric Ellis]

This sounds like an installation problem somewhere, I suggest that you contact MathWorks support to resolve this.

[Tyler Warner]

Thanks for the assistance. I ran "restoredefaultpath" and the parpool runs on my PC! However, it doesn't work with my server on Windows server 2012. The following error occurs:

>> restoredefaultpath
Warning: Duplicate directory name: C:\Program
Files\MATLAB\R2016a\toolbox\stateflow\stateflow 
> In restoredefaultpath (line 52)

Answer 1

[Kanikesh Yedla]

I faced same problem, but finally find the solotuion

--> Enter matlab command window 'prefdir' --> You can see C:\Users\xxx\AppData\Roaming\MathWorks\MATLAB\R2015a --> Delete local_cluster_jobs and restart matlab --> I hope this will fix the issue.

-- Kanikesh Yedla

[NLD]

Thanks for your answer. I was trying to run matlab jobs on our server at school, but it would give me this error. No one in the IT support could solve this error. You saved my life.

[Chifeng Liang]

Thanks for your perfect answer for this huge problem. I deleted the file, which you mentioned, and restarted the matlab program, and it work!

[Suvidha Tripathi]

I found no such file in the mentioned path. My problem still persists

[Walter Roberson]

Suvidha Tripathi: what results do you see for cluster validation?

[Suvidha Tripathi]

I found the file and deleted it after which I could successfuly start my parallel pool. I also did cluster validation and found everything ok. However, One problem still remains that I am unable to use my all GPUs for multi-gpu processing. I am going to raise a separate question for this.

Answer 2

[nikos skordilis]

Hi I have the same problem but local_cluster_jobs folder doesn't exist. Can anyone help me?

[Gabriel Limeira]

The "local_cluster_jobs" folder is one level up, in C:\Users\xxx\AppData\Roaming\MathWorks\ . Deleting it and restarting Matlab solves it.

[Yichen]

Same problem. And this solution works in 7/26/2017. Thanks.

[Longsheng Li]

You are the kind of person that we call superhero.

[Andrea Federico Loinger]

Thank you for this tip! Extremely helpful and easy to do! I confirm it is still valid!

[longlong leng]

Wow. Yes. You are amazing. Succeed.

[Jenni Stanley]

Awesome! This worked for me also!

Answer 3

[Kanikesh Yedla]

Usually it will fix if u delete "local_cluster_jobs" folder at C:\Users\xxx\AppData\Roaming\MathWorks\

[Niels Carl Hansen]

And btw, if you are on Linux, delete the directory ~user/.matlab/local_cluster_jobs

[Allister Mason]

thanks

Answer 4

[Christoph Schöne]

Hello Kanikesh Yedla,

thank you very much for this solution but I haven't got a file called "local_cluster_jobs". I've got a file called "parallel.settings" and deleted this one. But this does not fix the problem.

I think it's not a pure Matlab problem. I've reinstalled Matlab now two times and the error is still there. I've also updated the GPU Drivers (NVIDIA GEFORCE GTX960). but this doesn't solve the problem, too.

[Jien Cao]

I ran into the same issue just now and Kanikesh's solution worked - the missing detail is that the "local_cluster_jobs" folder is one level up, in C:\Users\xxx\AppData\Roaming\MathWorks\ . Deleting it and restarting Matlab fixed it for me.

[Bruno Remillard]

Yes it worked also for me. I just removed the .mat file and restarted Matlab

[Gabriel Limeira]

It worked for me too.

Answer 5

[muhammad irfan]

function dimg = mipbackwarddiff(img,direction)
% MIPFORWARDDIFF     Finite difference calculations 
%
%   DIMG = MIPBACWARDKDIFF(IMG,DIRECTION)
%
%  Calculates the central-difference for a given direction
%  IMG       : input image
%  DIRECTION : 'dx' or 'dy'
%  DIMG      : resultant image
%
%   See also MIPCENTRALDIFF MIPFORWARDDIFF MIPSECONDDERIV
%   MIPSECONDPARTIALDERIV

%   Omer Demirkaya, Musa Asyali, Prasana Shaoo, ... 9/1/06
%   Medical Image Processing Toolbox

img = padarray(img,[1 1],'symmetric','both');
[row,col]=size(img);
dimg = zeros(row,col);
switch (direction)
    case 'dx',
        dimg(:,2:col) = img(:,2:col)-img(:,1:col-1);
    case 'dy',
        dimg(2:row,:) = img(2:row,:)-img(1:row-1,:);
    otherwise, disp('Direction is unknown');
end;
dimg = dimg(2:end-1,2:end-1);

when i will run this program i face the error like this

(Error using mipbackwarddiff (line 17)

Not enough input arguments.)

kindly someone help me to solve this

[Walter Roberson]

You cannot run that code just be clicking the green Run triangle. You need either call it from other code or else go down to the command window, set up whatever variables you need, and then invoke it at the command window. For example,

img = imread('cameraman.tif');
dimg = mipbackwarddiff(img, 'dx');