On Wed, Jun 15, 2011 at 1:52 AM, Philip Martin <philip.mar...@wandisco.com> wrote: > Philip Martin <philip.mar...@wandisco.com> writes: > >> Johan Corveleyn <jcor...@gmail.com> writes: >> >>> Also, has anyone tested this on an NFS-working copy? Or CIFS? >> >> Working copy on NFS, Linux client and server. >> >> Checkout using 1.6: >> Elapsed: 96s CPU: 16s >> >> Upgrade using 1.7: >> Elapsed: 147s CPU: 21s >> >> Checkout using 1.7 >> Elapsed: 216s CPU: 26s >> >> So upgrade is faster than checkout, but only because checkout is slow. > > That's using the sync option on the NFS server. Using async > > Checkout using 1.6: > Elapsed: 73s CPU: 16s > > Upgrade using 1.7: > Elapsed: 113s CPU: 20s > > Checkout using 1.7 > Elapsed: 180s CPU: 26s
Wow, that looks pretty bad. Especially the checkout time is a cause for concern IMHO. That may point to a very significant performance regression accross the board when working with NFS working copies. Philip, any chance you can perform a complete test run (comparing 1.6 to trunk) of Mark Phippard's perf testsuite? Or Neels' one? -- Johan