RE: Performance of mid-sized tables?

Joe: 1. You may want put this table on seperate disk with its own disk controller. 2. You can create a set of tables to cache the frash data and put them back into the big tables when they getting old. All your program will search the cache table first. 3. Try to split a complex quary to a few simple one and store the temp result into temp tables. I am not sure SE5.02 can store data on raw disk or not, raw disk will be faster than unix files. Hope this helps. William Zhu ---------- From: informix-list-owner To: informix-list Subject: Performance of mid-sized tables? Date: Tuesday, August 06, 1996 6:07PM We're using SE5.02 and our most-heavily read tables are getting up to about 100,000 records. We've noticed that the seek and reads on them is degrading. Rather than splitting them to 50,000 tables, does anyone have an idea of how to improve performance? On a similar issue, we are running on a HP-UX K200 system (192meg ram) and recently added the second processor. We noticed that the system got slower rather than faster. Does anyone know how to split the processing up among the processors? Thanks, Joseph Kormann - programmer jkormann@rfbd.org