On 07/14/2010 08:09 PM, Steve French wrote: > On Wed, Jul 14, 2010 at 12:41 PM, Scott Lovenberg > <[email protected]> wrote: >> On 7/5/2010 8:41 AM, Suresh Jayaraman wrote: >>> >>> This patchset is a second try at adding persistent, local caching facility >>> for >>> CIFS using the FS-Cache interface. >>> >>> >> >> Just wondering, have you bench marked this at all? �I'd be interested to see >> how this compares (performance and scaling) to an oplock-centric design. >>
Yes, I have done a few performance benchmarks with the cifs client (and not SMB2) and I'll post them early nextweek when I'm back (as I'm travelling now). However, I have never done scalability tests (not sure whether there is a way to simulate a number of cifs clients). >> I'd hazard a guess that with pipelining support in SMB2 the performance will >> be even better since you can have a hot cache and more requests in flight. > > Yes - very plausibly > I have not tried the new SMB2 client. But, it seems the pipelining support, Oplocks (only Level II kind) could help improve performance. Thanks, -- Suresh Jayaraman -- To unsubscribe from this list: send the line "unsubscribe linux-cifs" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
