I'm trying to test parallel symbolic matrix inversion on cluster with gridMathematica 7.0 for Linux x86 (64-bit) (February 18, 2009). I use next commands:

LaunchKernels[RemoteMachine["node1",2]]  //Out: {KernelObject[11, node1],  KernelObject[12, node1]}
data=Import["/home/myhome/500.txt", "Table"]; //square 500x500 matrix with integers
m = ToExpression[data];
Parallelize[a = Inverse[m]];

But it issue next error:

a = Inverse[m] cannot be parallelized; proceeding with sequential

What's wrong with my commands? How can I evaluate symbolic matrix inversion on cluster with multiple nodes?

  • 1
    $\begingroup$ Are you really, really sure that you want to compute the inverse symbolically? What do you expect to see afterwards? In general, what you attempt is a very, very,very bad thing to do (both symbolically and numerically). That's why nobody would ever implement a parallel matrix inversions... $\endgroup$ – Henrik Schumacher Dec 3 '17 at 22:34
  • $\begingroup$ The message says it all. Symbolic Inverse has no code to support computing it by parallelized kernels. $\endgroup$ – Daniel Lichtblau Dec 7 '17 at 4:53

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.