Weave Insert Time

[John A Meinel]

John A Meinel wrote:
> Gustavo Niemeyer wrote:
>
>>>You can see the resultant graph here:
>>>http://bzr.arbash-meinel.com/other/RevisionInsertTime.png
>>
>>
>>Nice graph!
>>
>>
>>
>>>Basically, it looks like the time to insert one revision is very close
>>>to O(N) where N is the number of revisions already in the weave. I think
>>>that makes it O(N^2) when you are creating the weave for the first time.
>>
>>
>>One thing I'm wondering is: would it be O(N), where N is the number of
>>revisions, or O(N), where N is the number of lines in the weave?
>>In other words, I'm curious about what would happen with that graph if
>>you changed the size of the log messages in the revision, for instance.
>>
>
>
> I wrote a new plugin, available here:
> http://bzr.arbash-meinel.com/plugins/timeinventory-weave/
>

I have some plots. Available here:
http://bzr.arbash-meinel.com/other/timing/

The .csv is the raw data file.

Right now, just for my bzr-sax branch, where each attribute gets its own
line.

The most interenting graph to me is the weavelines.png
Which shows how the number of line starts to explode around 1500 or so.
I think that is mostly because newformat started, and in general bzr got
a bunch more committers at that time.

John
=:->

> It performs a whole bunch of operations on the inventory of the current
> weavefile, and creates a comma separated file with all the times and
> lengths, and any other neat thing that I could think of.
>
> Be forewarned, it is pretty slow (one of the things I optionally do is
> to extract *all* of the entries in the inventory).
>
> I'm letting it run right now, so I don't know all the information about
> how long different things are going to be.
>
> But I thought I would let people know about the plugin, so they can play
> around with their own branches.
>
> Actually, it could easily be adapted to take a weave file, and produce a
> second test weave file, and perform similar operations. So that we could
> perform the test against an inventory or a regular text file, etc.
>
> But I thought inventories were the most interesting, and it also gives
> me the opportunity to check on some inventory read/write times.
>
> John
> =:->

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 253 bytes
Desc: OpenPGP digital signature
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20051005/1643c59f/attachment.pgp