Twenty years or so ago, it seemed to be standard practice in the Defence Industry to write Fortran that way. For some reason, Named Common Blocks were thought to be bad practice, so everything was passed by formal parameters. Hundreds of them. Even if they weren't used in that routine;( think of a tree structure; to get info from one twig to another the data has to go all the way up to the trunk and back down again). More than one large simulation that I took over was spending almost all the execution time (overnight runs) pushing data onto the stack and pulling it off again, and was speeded up by orders of magnitude by use of Named Common.
Jeremy ----- Original Message ----- From: "Marcel Kilgus" <[EMAIL PROTECTED]> To: "ql-users" <[EMAIL PROTECTED]> Sent: Wednesday, February 05, 2003 3:08 PM Subject: Re: [ql-users] Display_cde GD2 update > > Wolfgang Lenerz wrote: > > 358 parameters? > > To a basic procedure? > > I'd be MOST interested in knowing what that procedure does... > > Nothing exciting really. Just seems that instead of using global > variables everything is passed as parameters (>100 alone are for > different key codes of the menu selection). Nice also the ~370 REMark > lines trailing the procedure that explain all parameters ;-) Which > also partly explains the size of over 13300 lines. > > Marcel >
