tag:blogger.com,1999:blog-5309604992043123290.post811788108543073079..comments2023-06-13T13:04:02.796+02:00Comments on berxblog: Instance parameters derived from cpu_countMartin Bergerhttp://www.blogger.com/profile/16504572924713610305noreply@blogger.comBlogger1125tag:blogger.com,1999:blog-5309604992043123290.post-65904968967104312002009-11-06T21:23:38.532+01:002009-11-06T21:23:38.532+01:00Martin, thanks for putting these up! Very interest...Martin, thanks for putting these up! Very interesting indeed. For instance, the log_buffer increases exponentially? Wow. Wonder what the "good reason" for that is. Also, interesting to note that for every 1.8 blocks the db_cache decreases, the shared_pool goes up. Why would db_cache go down at all? Why does shared_pool go up?<br /><br />We have observed similar behavior with the various writers, especially when you attempt to shut down your database. *grin* It takes a lot longer.<br /><br />One last observation, it is interesting how certain parameters quickly hit a ceiling (or a floor), and that any further changes in CPU_COUNT is completely irrelevant.<br /><br />Thanks again for putting this up; good stuff. I applaud your curiosity. *grin*Charles Schultzhttps://www.blogger.com/profile/07973399674184183130noreply@blogger.com