On Thu, Mar 04, 2004 at 06:57:37PM -0800, Kelsey Cummings wrote: > > As our testing has shown, SA Bayes' engine has brutal I/O requirements > using DB_File. Perhaps SQL could be far more efficient? Does anyone > running SQL Bayes have a comparison of I/O profiles between SQL and > DB_File? >
I've achieved slightly better (by a few %) results using MySQL on my test box compared to DB_File. I tend to believe that a properly tuned SQL server on good hardware would be much better than DB_File. I'd like to come up with some sort of comprehensive Bayes benchmark test (ie learn, scan, forget, etc) that everyone agrees tests/stresses the bayes code and then run it against several different methods (DB_File, QDBM, TDB, SQL, etc) to see how well each works. Michael
