From: "Jon Smirl" <jonsmirl@gmail.com>
To: "Shawn O. Pearce" <spearce@spearce.org>
Cc: "Martin Koegler" <mkoegler@auto.tuwien.ac.at>,
"Git Mailing List" <git@vger.kernel.org>
Subject: Re: performance on repack
Date: Tue, 14 Aug 2007 00:10:31 -0400 [thread overview]
Message-ID: <9e4733910708132110u6cdf5e6bg10417317c70b82f1@mail.gmail.com> (raw)
In-Reply-To: <20070814031236.GC27913@spearce.org>
On 8/13/07, Shawn O. Pearce <spearce@spearce.org> wrote:
> > On 32b there's windowing code for accessing the packfile since we can
> > run out of address space, does this code get turned off for 64b?
>
> The windowing code you are talking about defaults as follows:
>
> Parameter 32b 64b
> -----------------------------------------
> core.packedGitWindowSize 32M 1G
> core.packedGitLimit 256M 8G
>
> So I doubt you are having issues with the windowing code on a 64b
> system, unless your repository is just *huge*. I did not think that
> anyone had a Git repository that exceeded 8G, though the window
> size of 1G might be a tad too small if there are many packfiles
> and they are each larger than 1G.
Why use windows on 64b? Default core.packedGitWindowSize equal to
core.packedGitLimit
I haven't measured it but I suspect the OS calls for moving the
windows are are quite slow on a relative basis since they have to
rewrite a bunch of page tables. Why is the window so small on 32b? I
thought we were up to about a 1GB packfile before running out of
address space with Mozilla. Shouldn't the window simply be set as
large as possible on 32b, this size being a function of the available
address space, not the amount of physical memory?
> > > * On the other hand, we could run all try_delta operations for one object
> > > parallel. This way, we would need not very much more memory, but
> > > require more synchronization (and more complex code).
> >
> > This solution was my first thought too. Use the main thread to get
> > everything needed for the object into RAM, then multi-thread the
> > compute bound, in-memory delta search operation. Shared CPU caches
> > might make this very fast.
>
> I have been thinking about doing this, especially now that the
> default window size is much larger. I think the default is up as
> high as 50, which means we'd keep that shiny new UltraSPARC T2 busy.
> Not that I have one... so anyone from Sun is welcome to send me
> one if they want. ;-)
>
> I'm not sure its that complex to run all try_delta calls of the
> current window in parallel. Might be a simple enough change that
> its actually worth the extra complexity, especially with these
> multi-core systems being so readily available. Repacking is the
> most CPU intensive operation Git performs, and the one that is also
> the easiest to make parallel.
>
> Maybe someone else will beat me to it, but if not I might give such
> a patch a shot in a few weeks.
I'll test it. But my assignment this week is to figure out how to
program the BestComm DMA engine in a PowerPC chip. It's oh so much fun
trying to figure out how to program these "engines" that ASIC
designers love to build and never fully document.
>
> --
> Shawn.
>
--
Jon Smirl
jonsmirl@gmail.com
next prev parent reply other threads:[~2007-08-14 4:10 UTC|newest]
Thread overview: 30+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-08-11 21:12 performance on repack Jon Smirl
2007-08-11 22:09 ` David Kastrup
2007-08-11 22:34 ` Linus Torvalds
2007-08-11 23:21 ` Jon Smirl
2007-08-12 10:33 ` Martin Koegler
2007-08-12 13:49 ` Jon Smirl
2007-08-14 3:12 ` Shawn O. Pearce
2007-08-14 4:10 ` Jon Smirl [this message]
2007-08-14 5:13 ` Shawn O. Pearce
2007-08-14 5:57 ` Jon Smirl
2007-08-14 14:52 ` Nicolas Pitre
2007-08-14 21:41 ` Nicolas Pitre
2007-08-15 1:20 ` Jon Smirl
2007-08-15 1:59 ` Nicolas Pitre
2007-08-15 5:32 ` Shawn O. Pearce
2007-08-15 15:08 ` Jon Smirl
2007-08-15 17:11 ` Martin Koegler
2007-08-15 18:38 ` Jon Smirl
2007-08-15 19:00 ` Nicolas Pitre
2007-08-15 19:42 ` Jon Smirl
2007-08-16 8:10 ` David Kastrup
2007-08-16 15:34 ` Nicolas Pitre
2007-08-16 16:13 ` Jon Smirl
2007-08-16 16:21 ` Nicolas Pitre
2007-08-15 21:05 ` Nicolas Pitre
2007-08-15 20:49 ` Nicolas Pitre
2007-08-30 4:27 ` Nicolas Pitre
2007-08-30 4:36 ` Nicolas Pitre
2007-08-30 16:17 ` Jon Smirl
2007-09-01 21:54 ` Jon Smirl
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
List information: http://vger.kernel.org/majordomo-info.html
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=9e4733910708132110u6cdf5e6bg10417317c70b82f1@mail.gmail.com \
--to=jonsmirl@gmail.com \
--cc=git@vger.kernel.org \
--cc=mkoegler@auto.tuwien.ac.at \
--cc=spearce@spearce.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
Code repositories for project(s) associated with this public inbox
https://80x24.org/mirrors/git.git
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).