Re: RFC: New diff-delta.c implementation

From: Geert Bosch <bosch@adacore.com>
Date: Apr 24, 2006, 18:44 UTC
Message-ID: <A175169A-0875-4DEF-BE97-C556325EB6DB@adacore.com>
In-Reply-To: <Pine.LNX.4.64.0604232327500.3603@localhost.localdomain>

On Apr 24, 2006, at 01:27, Nicolas Pitre wrote:

Show 5 quoted lines

> But here comes the sad part.  Even after simplifying the code as  
> much as
> I could, performance is still significantly worse than the current
> diff-delta.c code.  Repacking again the same Linux kernel repository
> with the current code:

Changing the level parameter in the call to init_idx to 0 gives a significant speedup. After that, by far the most time is spent computing hashes. I have some ideas of cutting down on that for this test case.

> The final pack is smaller with your code but not significantly:
> 117867049 bytes vs 118824550 bytes with the current code, i.e. less  
> than
> 1% difference.

I'm doing tests on the git.git repository now, and even though I see similar performance, I noted that the new algorithm packs more files, so the pack size is not the only thing to look at. Still, it remains the case that when you compare two files with just one or two changes, the simplest algorithm is still good enough.

It will be very interesting how things will work out when comparing 10 files at a time. Then the extra cost of building the index isn't that significant, and the higher quality of the index may then pay off.

   -Geert