Paul Burba <ptbu...@gmail.com> writes: > No surprise that your checkouts are faster than mine given you are > using a local mirror. What's puzzling me is why my upgrades are so > much slower than yours. > > Running an upgrade of a trunk WC on my machine under xperf takes > 00:03:46.351 elapsed and 11.44s CPU time using my primary drive (320 > GB, SATA-II, 7200 rpm, 16 MB Cache, NTFS). Subversion spends 50s > total disk service time (46.8s of that is read service time).
Does "total disk service time" represent all the time waiting for disk IO? 11s CPU plus 50s disk still leaves 165s unaccounted. Upgrade prints a notification line for each directory, is there a significant delay before the first line, or after the last line, or is the delay spread among the lines? If it is a Subversion problem rather than your machine there are two areas that may be worth investigating (but it's hard to say when most of the time is unaccounted). The property migration currently invokes multiple transactions per-directory. Moving to a single transaction per-directory will probably help. Upgrade currently copies all the text-bases, I did experiment with creating workqueue items to move them instead but it wasn't any faster on my Linux box. However it may help on Windows. -- Philip
