mohan@ihuxf.UUCP (Palat) (08/23/85)
Has anyone noticed that the time to invoke ingres is much more the first time after a compilation (of an equel program) than the subsequent ones? That is, if the same compiled equel program is executed two or more times, the time taken to log in to ingres is longer the first time than the subsequent executions. Any possible reasons for this? By the way, the version of ingres used was AT&T INGRES. Mohan Palat
anton@ucbvax.ARPA (Jeff Anton) (08/24/85)
In article <2671@ihuxf.UUCP> mohan@ihuxf.UUCP (Palat) writes: > Has anyone noticed that the time to invoke ingres > is much more the first time after a compilation (of an equel > program) than the subsequent ones? That is, if the > same compiled equel program is executed two or more > times, the time taken to log in to ingres is longer the > first time than the subsequent executions. Any possible > reasons for this? By the way, the version of ingres used > was AT&T INGRES. RTI INGRES has a method of storeing query plans in equel programs. I'm not sure but perhaps the database is conditioned or these plans are stored durring execution. I've not heard of AT&T INGRES. Is it not an RTI port? -- C knows no bounds. Jeff Anton U.C.Berkeley ucbvax!anton anton@BERKELEY.EDU