Re: db block size, too big wasting buffer?

sg <s4v4g3_at_europe.com> wrote in message news:<3D178B57.3010408_at_europe.com>...
> Has anyone benchmark this? Or this just theory you peeps are talking about?

I've been searching for benchmarks for years. So far only two cases:

http://www.linuxjournal.com/article.php?sid=5840&mode=thread&order=0 (particularly read the section DB6: 8K Block) Here the author claimed 8k block size improved performance over 4k on Linux, where ext2 filesystem block size is 4k.

http://groups.google.com/groups?hl=en&safe=off&th=4e8ae10f1c3c4d49,2&seekm=3b2fba18%40news.iprimus.com.au Here Nuto Souto reported performance increase for batch jobs when db_block_size is changed from 2k to 4k. Not a benchmark but is a good comparison to remember.

With so much time wasted debating, one could do a very thorough benchmark convincing everybody. But nobody has done so. (Me included!)

Yong Huang Received on Tue Jun 25 2002 - 12:04:25 CDT