Date: Thu, 8 May 2014 09:49:02 +0200 From: Borja Marcos <borjam@sarenet.es> To: Kenneth D. Merry <ken@freebsd.org> Cc: freebsd-scsi@freebsd.org Subject: Re: Testing new mpr driver Message-ID: <4DB83981-0D4D-4484-BC89-4ED8C02DCD0F@sarenet.es> In-Reply-To: <20140507184557.GA80243@nargothrond.kdm.org> References: <8A41AB90-AC2F-4200-91D6-3D3CF9E8A835@sarenet.es> <20140507184557.GA80243@nargothrond.kdm.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On May 7, 2014, at 8:45 PM, Kenneth D. Merry wrote: > That's hard to say. If you're using a 6Gb expander, you would have = half of > the available SAS bandwidth if you only connected four lanes from the > controller to the expander instead of 8. If you somehow have a 12Gb > expander (it isn't obvious from the model number above what the = expander > speed is), then you would have the same amount of bandwidth. Anyway, as far as I understand (SAS expanders perform link switching, = right?) the actual speed will be limited by the end to end speed. As the disks I am using are SATA, not SAS, each = lane would be anyway working at=20 6 Gbps instead of 12.=20 > One thing that could be happening is you may have lower latency = through the > new 12Gb controller. As I saw it supports something called "fastpath" (I must read something, = I am a bit outdated in these matters) I imagined that it=20 might be a more efficient transfer method which, despite working at 6 = Gbps, might explain a gain in performance. > By the way, if you run with INVARIANTS enabled, you may run into some > issues (i.e. a panic) on reboot until we merge r265485 to stable/10. I am running stable/10, and I'm not running INVARIANTS. Anyway I will = be tracking stable/10 closely, thank you! Great to have LSI so seriously involved! Borja.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4DB83981-0D4D-4484-BC89-4ED8C02DCD0F>