I also would benefit from parallelism in the kcryptd support. I replied to a previous thread on this. My situation is outlined in a bug report to Ubuntu: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/246413 Although I managed to wipe that entire array :( by being an ignorant retard and lost 4TB of data... I will be rebuilding it eventually. Before starting I did not realize that the only way to get multiple kcryptd processes was to encrypt each device separately, or else I would have considered that. However, it would be a huge pain to have to open 7 LUKS devices just to bring my array online (especially when using the recommended 30-40 char passphrase). Each drive can do ~100MB/s, so even if this weren't a raid array I believe that I would see a performance increase from parallelism even just encrypting one drive (as Micheal mentioned). Someone mentioned a specialized hardware encryption (maybe on the last thread)... What options are there for hardware assisted encryption? In short, I think more people would benefit from this functionality than Arno, and others, believe. Regards, -Clay On Wed, Sep 17, 2008 at 3:25 AM, Michael Gebetsroither <gebi@xxxxxxxxxxxxxx>wrote: > * Arno Wagner <arno@xxxxxxxxxxx> wrote: > > I have about 45% CPU on a very low end AMD Sempron(tm) > > Processor LE-1250 for 30MB/s. You probably should have > > bought faster cores instead of more. > > 30MB/s? > You know that even a single moderm SATA disk does 100MB/s streaming? > The problem is, that about 0.1% of the users have CPU's where one core > is capable of >100MB/s of even AES 128. > > The _real_ problem behind all these is much bigger. For the majority of > the users the single-threaded performance will stagnate (or even > decrease) in the next years > > > Incidentially, I believe your request has about zero prospect > > of being sucessful. It is a lot of effort for basically > > very few people having any gain. I would not do it unless fully > > paid, but thet _wpuld_ be expensive. > > As the fix with the current situation is quite simple (put the raid on > top of dm-crypt) it's not the biggest problem, yes. > But as stated above, moderm disks are faster than one core could encrypt > with AES128. > And the situation only gets worse. > > cu, > michael > -- > It's already too late! > > > --------------------------------------------------------------------- > dm-crypt mailing list - http://www.saout.de/misc/dm-crypt/ > To unsubscribe, e-mail: dm-crypt-unsubscribe@xxxxxxxx > For additional commands, e-mail: dm-crypt-help@xxxxxxxx > >
