Re: [rfc] Near-constant time directory index for Ext2

[Linus Torvalds]

On Thu, 22 Feb 2001, Daniel Phillips wrote:
> 
> In the first heat of hash races - creating 20,000 files in one directory
> - dentry::hash lost out to my original hack::dx_hash, causing a high
> percentage of leaf blocks to remain exactly half full and slowing down
> the whole thing by about 5%.  (This was under uml - I haven't tried it
> native yet but I expect the results to be similar.)
> 
> 	  Contender			Result
> 	  =========			======
> 	dentry::hash		Average fullness = 2352 (57%)
> 	hack::dx_hash		Average fullness = 2758 (67%)
> 
> This suggests that dentry::hash is producing distinctly non-dispersed
> results and needs to be subjected to further scrutiny.  I'll run the
> next heat of hash races tomorrow, probably with R5, and CRC32 too if I
> have time.

I'd love to hear the results from R5, as that seems to be the reiserfs
favourite, and I'm trying it out in 2.4.2 because it was so easy to plug
in..

		Linus