Hi all,
gotta do some computations involving a really big matrix, and need to
save each row in a different file (matrix is too big to fit in
memory). SInce every row can be computed independently I have tried to
parallelize the computation. First, this is my original computation of
the matrix (for a group G):
def computeKillingbyrows(G):
print("Studying "+str(G))
print("computing normalizers")
normalizer_sizes = dict([(g, G.normalizer(g).order()) for g in G])
l = sorted(normalizer_sizes)
n = len(l)
print("computing Killing form")
for x in range(n):
compute_row(x)
row = [normalizer_sizes[l[x]*l[y]] for y in xrange(n)]
filename = str(G)+"_row_"+str(x)
save(row, filename)
print("Row "+str(x)+" saved!")
one can easily check that this works for a small group (say A4). Now,
this is what I try as a parallel version (compute and save each row
independently)
cores = 2 # just for testing, will adjust later
def computeKillingbyrows(G):
print("Studying "+str(G))
print("computing normalizers")
normalizer_sizes = dict([(g, G.normalizer(g).order()) for g in G])
l = sorted(normalizer_sizes)
n = len(l)
@parallel(cores)
def compute_row(x):
row = [normalizer_sizes[l[x]*l[y]] for y in xrange(n)]
filename = str(G)+"_row_"+str(x)
save(row, filename)
print("Row "+str(x)+" saved!")
print("computing Killing form")
compute_row(range(n))
the latest gets to print down the "computing Killing form" print, and
exits without any error and without saving any files.
Is it possible to do the above with @parallel? I have tried looking
for documentation for that decorator, but the examples in the source
don't try to do anything similar.
Cheers
J
--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org
