Lifting Surface Program (LSP)

I've attached a small lifting surface program that some might find useful.

It would be handy if someone could do some timing runs on their computers to inform others of the execution times on a variety of computers.
A good example would be to run the program for rectangular planforms, 32 chordwise panels, and for 10 aspect ratios. That should only take a minute or two.

There is a small compendium of results that I posted on another thread that you should also download:
http://www.boatdesign.net/forums/sailboats/myth-aspect-ratio-36836-7.html#post446638

If there is interest, I will add more planforms and post an updated compendium in this thread.

And please remember, this is not a wing-design program. There are much better and far more versatile programs available if you need that capability.

Leo.

 

on a Toshiba laptop with Pentium T2330 1.6Ghz 2Gb and external screen the program ran about 5 min.

edit: the "during run" screenshot is off a bit due to my cut and paste error

 

Thanks, John.

I have nearly finished the manual for the more general (i.e. non-demo) version in which users can input their own curves defining the leading-edge and trailing-edge. This version will also be a little more accurate for the same number of panels.


Leo.

 

Using the same Intel T2330 dual core 1.6Ghz laptop:

Running this again after updating and restarting my computer, then putting a clock on the screen I got 4:50 run time for one instance.

Running two instances gave close to the same amount of time. Both finished 5 min after starting them.

 

Running on an AMD 7750 dual core 2.6Ghz cpu 32 bit Windows XP desktop the program finishes in 3:40 for two simultaneous instances.

A single instance finished in 3:38

 

3.0Ghz pentium 4 single core HT, 32bit windows

single instance 3:40

double instance 6:30

(yeah I can paddle a canoe 50+ miles in a day and still be this much of a nerd....)

 

That's unusual isn't it? I thought the processes would each take about the same, or does this machine have little memory?

Nerd schmerd. I wrote the code while watching B-grade scifi shows and eating cold pizza for breakfast, because that's the way I like it

 

On the dual core tests the 2 instance finished in the same amount of time as the single instance. All that happened is that the second core became active and they processed in parallel.

The P4 is a single core processor so running a single instance of a program like this can take up all the processing power. Adding a second process forces them to share CPU time.

The P4 does do "Hyperthreading" which means it has 2 pipelines to the CPU that can be loading the next step while the current one is being completed.
It's something of a counterfit dual cpu.


BTW I do wish that Godzilla would continue to run in the background like this program does.

 

Ah yes, now I remember how they work.

The graphics version won't run in the background. My text versions produce no output while running and can be run in the background, but they require the number of evaluations to be specified in advance.

Thanks again,
Leo.