On Du, 06 iul 14, 23:30:28, Bob Proulx wrote: > Andrei POPESCU wrote: > > Bob Proulx wrote: > > > Of course now with SSDs that standard thinking needs to be thought out > > > again. I haven't seen any benchmark data for full SSDs. I imagine > > > that it will have much flatter performance curves up to very full on > > > an SSD. It would super awesome if someone has already done this > > > performance benchmarking and would post a link to it so that we could > > > all learn from it. > > > > As far as I understand SSDs can drop performance significantly when > > full, but for entirely different reasons. Probably a good start would be > > this Wikipedia article: > > https://en.wikipedia.org/wiki/Write_amplification , especially the > > section on over-provisioning. > > Over-provisioning is the reserved space that the vendor keeps to so > that the physical layer of the storage device never actually becomes > very full. The more the vendor does over-provisioning the less full > the drive can get. Even if the logical layer appears full. The drive > needs big blocks of free space in order to be able to operate since it > will erase in large blocks. And of course that reduces the impact of > not using trim. This is why I don't think an SSD would degrade in > performance very much at all when the logical layer of the file system > is very full. Well, my point was that the reserved blocks would be useful for SSDs as well, even if for entirely different reasons ;) Kind regards, Andrei -- http://wiki.debian.org/FAQsFromDebianUser Offtopic discussions among Debian users and developers: http://lists.alioth.debian.org/mailman/listinfo/d-community-offtopic http://nuvreauspam.ro/gpg-transition.txt
Attachment:
signature.asc
Description: Digital signature