|
One other thing:
If you decide to go with straight BTree, you might
want to (re?) read:
On the Advanced Optimistic Algorithms for Concurrent B-Tree
http://citeseer.nj.nec.com/cache/papers/cs/22880/http:zSzzSzwww.xrce.xerox.comzSzpeoplezSzchidlovskiizSz.zSzpublicationzSzbiwit95.pdf/on-the-advanced-optimistic.pdf and maybe even (after all, it may also be possible
to easily extend the curent AVL implementation as well):
Concurrent rebalancing of AVL Trees - A fine grained approach
http://citeseer.nj.nec.com/cache/papers/cs/171/http:zSzzSzwww-lsi.upc.eszSzdeptzSztechrepszSzpszSzR97-10.pdf/concurrent-rebalancing-of-avl.pdf Both of these papers seem to claim some advantages over B-link, although I have not read them for a while, nor did I take the time to really think them through on the first read to see if I agreed at some gut level enough to experiment ot try to determine formally if they have merit. ----- Original Message -----
From: Campbell Boucher-Burnet
Sent: Tuesday, April 23, 2002 7:20 AM
Subject: Re: fine grain threading Hi Karl and Fred.
...
|
