# Profile for dataset preparation phase.
#
# Large 8MiB sequential writes using 64 threads per host and using direct IO to
# bypass client cache.

block=8m
direct=1
iodepth=1
threads=64
write=1
