48amd64-4pa2n6c-sparse.test

# 48-core AMD Magny-Cours, initially interesting because:
# - 2 NUMA nodes inside each socket
# - non-multiple of 32 kernel cpumaps (NR_CPUS=48)
# Later modified to simulate sparse numa node OS indexes.
# The distances matrix was also modified to become asymmetric (to verify we don't mix rows/columns)
# but this change was lost somewhere in the history.
source:  48amd64-4pa2n6c-sparse.tar.bz2
target:  48amd64-4pa2n6c-sparse.console