I'm parallelizing some computation on hierarchical data, and would
find it natural to do something like this (where a call to parLapply
is embedded in outer call to parLapply):
cl <- makeCluster(rep.int('localhost', 5),
type='SOCK')
clusterExport(cl, 'cl')
parLapply(cl, 1:5, function(i) parLapply(cl, 1:5, function(j) i * j))
Snow executes without error, but produces NULL; is embedded
parallelization, in principle, possible?
