Dear Ernst, elodw <[email protected]> writes: > i want to start with pycuda and had a very beginner question: > > ok, the header lines are clear, > same to_gpu statement and the get statement, > thats it. > > What is the problem defined in Python: > > a=array[1000,100] > > for i in range(1000-1): > for j in range(i+1,1000): > calculate euclidean difference between a[i] and a[j] > > can i do this without using an extra kernel > if i need a kernel how looks it
I'm afraid, PyCUDA's arrays are not currently smart enough to do this, so you'll need to write your own kernel. I unfortunately can't write that for you, but you should be able to take a look at the CUDA docs and give it a shot. Hope that helps, Andreas
pgp3EsCJX5H7C.pgp
Description: PGP signature
_______________________________________________ PyCUDA mailing list [email protected] http://lists.tiker.net/listinfo/pycuda
