comp.lang.idl-pvwave archive: archive » Re: Randomu seed initialization

Home » Public Forums » archive » Re: Randomu seed initialization

Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend

Re: Randomu seed initialization [message #68605]

Fri, 13 November 2009 08:31

Conor
Messages: 138
Registered: February 2007

Senior Member

Thanks for the input guys:

Jeremy:

I've checked in my primary function, and the variable I pass is
uninitialized at first (obviously it becomes initialized after the
first call to randomu() and stays that way). I haven't specifically
checked that it is uninitialized in the subroutine, but the subroutine
is only about 5 lines long, and I'm certainly not initializing it. So
if it is initialized, then that's certainly an IDL bug (and it seems
unlikely to me).

Matt:

I checked this possiblity, by making a couple simple routines:

pro test
print,randomu(seed,1)
test2
end

pro test2
print,randomu(seed,1)
end

Run this and you will get two different answers out - i.e. the random
seed is not initialized to the same value more than once. Which is
why I don't understand what's going on with my routine: conceptually,
this is the simple case of what I'm doing in my routine. This is why
I really wish I knew what randomu() was doing. I'm wondering (and
there are kinda hints to this in the documentation) if there is a
hidden "global" seed, which is initialized by the first call to randomu
(), and used by all subsequent calls to randomu() that have an
uninitialized seed. Either that or the initialization of the seed is
quasi-random each time (which would be much better, IMO).

Chris:

Thanks for the suggestion, that's what I've done for now, and it fixes
it. I just wish I understood what is going on.

Paolo: I could definitely go for some better documentation!
Presumably if it was better documented I would be able to avoid
problems like this, and I would (perhaps) be better able to determine
what I was doing wrong, or if this is a bug. As near as I can
determine I'm using randomu properly (it's not exactly a complicated
function). To uninitialized variables go into two separate calls to
randomu(), and the same random numbers come out both times. Is that
supposed to happen ever?

Report message to a moderator

[Message index]

		Re: Randomu seed initialization By: Conor on Fri, 13 November 2009 08:31
		Re: Randomu seed initialization By: pgrigis on Thu, 12 November 2009 13:59
		Re: Randomu seed initialization By: Chris[6] on Thu, 12 November 2009 13:49
		Re: Randomu seed initialization By: Matt[2] on Thu, 12 November 2009 12:32
		Re: Randomu seed initialization By: Jeremy Bailin on Thu, 12 November 2009 11:31
		Re: Randomu seed initialization By: Conor on Thu, 12 November 2009 05:59
		Re: Randomu seed initialization By: penteado on Wed, 11 November 2009 12:22

Previous Topic:	how smooth surface
Next Topic:	Randomu() behavior - BAD!

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Wed Oct 08 19:12:53 PDT 2025

Total time taken to generate the page: 0.00601 seconds