Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!uunet!mcsun!tuvie!roland From: roland@tuvie (Inst.f.Allg.Elektrotechnik) Newsgroups: comp.sys.atari.st Subject: graf_mkstate Message-ID: <1160@tuvie> Date: 8 Mar 90 14:46:44 GMT Reply-To: roland@tuvie.UUCP (Inst.f.Allg.Elektrotechnik) Organization: Technical University of Vienna Lines: 48 A simple timing loop with graf_mkstate gives about 2500 microseconds per call. The Turbo-C program is appended to the end of this message. I use TOS 1.4 and the same result is obtained with Pascal. However it takes at least 0.1 seconds before graf_mkstate recognizes any button state change. So we have at least 40 calls to this AES routine before the change is detected. I cannot imagine any reason for this. If anybody knows, please give an explanation. Roland Schreier _ Technical University of Vienna Gusshausstrasse 27-29 |_) _ | _ _ _| +43 (222) 58801 3838 A-1040 Wien | \ (_) | (_| | | (_| SCHREIER@EAECL1.UNA.AT Oesterreich - Europe roland@tuvie.at /* ---------------------------------------------------------------- */ /* BWAIT Button Wait */ /* author Roland Schreier */ #include #include #include int main() { int mx, my, mstate, kstate ; char s[128] ; long count ; clock_t time ; appl_init() ; count = 0 ; time = clock() ; do { count++ ; graf_mkstate(& mx, & my, & mstate, & kstate) ; } while (mstate == 0) ; time = clock() - time ; sprintf(s, "[0][mstate = %#.4x |%li ticks |%li calls |%li microseconds ][ OK ]", mstate, time, count, (time * 1000000L) / (CLK_TCK * count)) ; form_alert(1, s) ; appl_exit() ; return 0 ; } /* main */ /* ---------------------------------------------------------------- */