2 small comments after I re-read my post (of course after hitting send again): > abbreviations). So if sessions start hitting a lot of never pinned prefetch > blocks, the x$kcbkpfs statistics start to indicate inefficient prefetching, > thus causing CBO not to favour it and prefer regular single block reads. What I meant here is that if lot's of never used prefetched blocks are *aged out*, then the prefetch statistics get worse. If those blocks actually get used, the stats will be ok. [snip] > Note that I'm only concluding my last claim, I've not done testing to see > whether extra branch blocks actually happen - but I don't see any other way I meant here "whether extra branch block READS actually happen"... Tanel. ---------------------------------------------------------------- Please see the official ORACLE-L FAQ: http://www.orafaq.com ---------------------------------------------------------------- To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx put 'unsubscribe' in the subject line. -- Archives are at //www.freelists.org/archives/oracle-l/ FAQ is at //www.freelists.org/help/fom-serve/cache/1.html -----------------------------------------------------------------