[OpenAFS-devel] good and bad performance of memory cache

[Jenkins, Steven]

Lyle will probably chime in here, but my recollection is that memcache
is using an array-based implementation of the cache instead of an hash;
thus searching is O(n) instead of O(log n). =20

Steven

-----Original Message-----
From: Edward Moy [mailto:emoy@apple.com]=20
Sent: Wednesday, February 05, 2003 3:44 PM
To: openafs-devel@openafs.org
Subject: [OpenAFS-devel] good and bad performance of memory cache

Since I was starting to investigate memory caching, I decided to run =20
iozone to really exercise the cache.  With disk caching on my =20
moderately slow (400MHz) test machine, the full iozone test suite took =20
nearly seven hours, with the largest files (going up to half a =20
gigabyte) really dragging down performance.  So much so that iozone =20
running over NFS2 actually runs 20 minutes faster than AFS.

I ran the same test with memory caching enabled (and even a larger =20
cache; 80 MB memory cache versus the 30 MB disk cache).  The whole test

suite now took nearly *twelve* hours.

In comparing results, speed is definitely faster for the memory cache =20
when the file size is smaller than the cache size.  But when the file =20
size is larger than the memory cache, performance is often worse (in =20
some cases more that twice as bad) as the disk cache.

Anyone have any ideas about why this is so?
------------------------------------------------------------------------

--
Edward Moy
Apple Computer, Inc.
emoy@apple.com

(This message is from me as a reader of this list, and not a statement
from Apple.)

_______________________________________________
OpenAFS-devel mailing list
OpenAFS-devel@openafs.org
https://lists.openafs.org/mailman/listinfo/openafs-devel