... Pitre wrote: >> In the mean time you might ... said, such a wholesale repack is not >> something that most users have to do anyway. > > Is there an alternative to "git ... be made permanent; then to repack really really everything you'd ...

git - Paolo Bonzini - Dec 13 2007 - 11:32

... pack could be made permanent; then > to repack really really everything you'd need "git repack -a -a -d". Actually there is something like this, as seen from the ... this list: send the line "unsubscribe git" in the body of a message to majordomo ...

git - Paolo Bonzini - Dec 13 2007 - 12:29

... Pitre wrote: >>> In the mean time you might ... said, such a wholesale repack is not >>> something that most users ... be made permanent; then > to repack really really everything you'd ... send the line "unsubscribe git" in the body of a message ...

git - Johannes Sixt - Dec 13 2007 - 12:39

... a pack that is 7 times smaller. OK, here's something else for ... with 1GB of ram using: git repack -a -f -d --window=250 -- ... reports a ~700m virt and ~500m res without hitting swap at all. It ... : send the line "unsubscribe git" in the body of a message ...

git - Nicolas Pitre - Dec 11 2007 - 11:00

... Nicolas Pitre wrote: > In the mean time you might have ... Linus said, such a wholesale repack is not > something that most users have to do anyway. Is there an alternative to "git ... : send the line "unsubscribe git" in the body of a message ...

git - Nguyen Thai Ngoc Duy - Dec 13 2007 - 09:32

... code and not the > > >> repack code. And it must have something to do with ... bloat with or without THREADED_DELTA_SEARCH defined? > > > > Something else seems to be wrong. > ... using multiple threads, the list is broken in parts for each thread, and ...

git - Nicolas Pitre - Dec 9 2007 - 21:07

... >> 3GB with 2GB of mmaps. > > > > Which is quite reasonable, even if the same issue ... the pack access code and not the > > repack code. And it must have something to do with ... and because old objects are typically in > > a deeper delta chain, then this ...

git - Nicolas Pitre - Dec 9 2007 - 22:49

... pack greatly aggravates the problem. That is really interesting though. > Starting with a 2GB ... and not the repack code. And it must have something to do with ... be wrong with the delta cache in sha1_file.c somehow. Nicolas - To unsubscribe ...

git - Nicolas Pitre - Dec 7 2007 - 21:46

... pack greatly aggravates the problem. > > That is really interesting though. > > > Starting with a 2GB ... code and not the > repack code. And it must have something to do with ... , and because old objects are typically in > a deeper delta chain, then this ...

git - Jon Smirl - Dec 7 2007 - 22:04

... pack greatly aggravates the problem. > > > > That is really interesting though. > > > > > Starting with a ... and not the > > repack code. And it must have something to do with ... and because old objects are typically in > > a deeper delta chain, then this ...

git - Nicolas Pitre - Dec 7 2007 - 22:28

... not the > repack code. And it must have something to do ... explain the logarithmic slowdown. > > So something must be wrong with the ... delta cache in sha1_file.c somehow. All I have is a ... second and a corresponding increase in process size). Didn't keep ...

git - Harvey Harrison - Dec 7 2007 - 23:44

... to get the default pack: $ git repack -a -f -d # first measurement with ... (210major+274887minor)pagefaults 0swaps This is with pack.threads=2 on a ... Virtual memory allocation never reached 700MB in both cases either. Nicolas - To unsubscribe ...

git - Nicolas Pitre - Dec 10 2007 - 15:56

... pack that is 7 times smaller. > > OK, here's something else ... 1GB > of ram using: > > git repack -a -f -d --window=250 ... hitting swap at all. > It is only at 25% so far ... just get laid down sequentially in memory. Separating these allocation types ...

git - Jon Smirl - Dec 11 2007 - 11:36

... EST) > BUT. The point is that repacking the gcc repo using ... later 300MB pack. If you repack on the smaller pack file ... to expand more stuff internally in order to search the deltas ... objects blobs for delta search. In fact that behavior makes perfect ...

git - David Miller - Dec 11 2007 - 13:24

... EST) > > > BUT. The point is that repacking the gcc repo using ... later 300MB pack. > > If you repack on the smaller pack file ... to expand more stuff > internally in order to search the deltas ... or not the source pack is densely packed or not. Nicolas ...

git - Nicolas Pitre - Dec 11 2007 - 13:44

... you actually want to do is to just re-use already packed delta ... the huge difference > between the two repack runs that have different starting ... , as in the 2.1GB pack case), do run "repack -f", ... eats ALL virtual memory) is really something that lowers git's ...

git - Andreas Ericsson - Dec 11 2007 - 16:34

... code and not the > repack code. And it must have something to do with the ... deltas > being replayed. And because the repack is attempting delta compression > roughly from ... search through more and more objects in memory. Memory consumption seem to be ...

git - Jon Smirl - Dec 7 2007 - 22:22

... >Now do the command again. > git repack -a -d -f --depth=1000 --window= ... --window-memory as well to keep repack from using nearly unbounded amounts of memory. Perhaps it is the larger ... list: send the line "unsubscribe git" in the body of a message to ...

git - David Brown - Dec 7 2007 - 23:37

... with 2GB of mmaps. > > Which is quite reasonable, even if the ... not the > repack code. And it must have something to do ... because old objects are typically in > a deeper delta chain, then ... to spot anything blatantly wrong in that area. Will need to ...

git - Junio C Hamano - Dec 8 2007 - 18:18

... with 2GB of mmaps. >> >> Which is quite reasonable, even if the ... not the >> repack code. And it must have something to do ... because old objects are typically in >> a deeper delta chain, then ... to spot anything > blatantly wrong in that area. Will need to ...

git - Junio C Hamano - Dec 9 2007 - 04:05