Open tanghaodong25 opened 5 years ago
If the pmem is mounted as 2M aligned, the pmem obj address can't be registered as rdma buffer. We need to mount pmem as 4K aligned, but the pmem write performance is worse than it with 2M aligned.
2M aligned (default) ndctl create-namespace -r region0 -f -e namespace0.0 -m devdax 4K aligned ndctl create-namespace -r region0 -f -e namespace0.0 -m devdax -a 4k
ndctl create-namespace -r region0 -f -e namespace0.0 -m devdax
ndctl create-namespace -r region0 -f -e namespace0.0 -m devdax -a 4k
we could set the alignment at pool level: https://github.com/pbalcer/nvml/blob/6f2136f379c9a44ad73315ab2075fe665e02bcf7/src/test/obj_ctl_alignment/obj_ctl_alignment.c
If the pmem is mounted as 2M aligned, the pmem obj address can't be registered as rdma buffer. We need to mount pmem as 4K aligned, but the pmem write performance is worse than it with 2M aligned.
2M aligned (default)
ndctl create-namespace -r region0 -f -e namespace0.0 -m devdax
4K alignedndctl create-namespace -r region0 -f -e namespace0.0 -m devdax -a 4k