|cino hilliard on Tue, 04 Apr 2006 23:09:28 +0200|
[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]
|Re: elapsed time and system command|
From: Joerg Arndt <email@example.com> To: firstname.lastname@example.org Subject: Re: elapsed time and system command Date: Tue, 4 Apr 2006 11:16:02 +0200
You now have compile-time workaround, so everything should be fine. You can always measure wall-clock time with /usr/bin/time You do (at least, also) want to know how many cycles where eaten. We might introduce a time format string that gives the possibility to see it all. As in % /usr/bin/time find 0.03user 0.14system 0:11.55elapsed 1%CPU Then we have it all. default(timeformat," %Uuser %Ssystem %Eelapsed %C%%CPU") <--= suggestion which at the moment would be default(timeformat,"time = %U ") as in ? factor(2^239-1) time = 23 ms.
Thanks for your effort to help. I am using the windiows binary in xp pro. (11:11) gp > \v GP/PARI CALCULATOR Version 2.2.12 (beta) i686 running cygwin (ix86 kernel) 32-bit version compiled: Jan 3 2006, gcc-3.4.1 (cygming special) (readline v5.0 enabled, extended help available) I could not implement the above. I changed my .gprc to prompt = "(%T) gp > "By using the arrow keys and return to retrieve the last input quickly with practice you can get a
good timeing. (11:15:42) gp > g(n) = for(x=1,n,write("testtime1.txt","12345678901")) (11:17:31) gp > g(100000) (11:18:30) gp > (11:18:46) gp > g(100000) (11:19:36) gp > The last two lines gives 50 sec as opposed to (11:19:36:19S) gp > ## *** last result computed in 11,360 ms. This is awkward. I changed it back. Back to the stop watch. Thanks again, Cino