Sunday, May 20, 2007

a promising start to the new design



First it looks like my estimation for the simulation was way off, in a good way. This animation was compiled from a 1000 frame/100 steps per frame simulation, and it only took 14 hours. Not bad. The sluggishness I was experiencing during construction did not carry over to the simulation. I believe this makes since because building the part, lots of rotations and view changes, may have taxed my video card and ram, and the speed of the simulation would depend more on the processor. That's my theory anyway. For the record, at a half gig, my computer could stand an upgrade. Also I am using an ATI Radeon 9500 video card, so not top of the line there either.

Maybe my dreams of falling asleep in the penumbra of a processor farm will have to wait, but I am sure I will need it someday, all rack mounted, dozens of twinkling LEDs tapping out messages from the future in pseudo-morse, radiator fins on the roof blazing the IR spectrum like some techno-creature in heat. Yes, someday.

I know this simulation was way too short to see if this design will ultimately work, but for right now I just want to make sure nothing started flying off the thing. Nothing did. You can see that the pinion starts to twist the casing too. Next I will anchor the casing at the corners (just imagine this casing is bonded to some monolithic substrate) and run a long simulation.

More to follow.

2 comments:

Guy said...

Hi there,
You might not have to acquire a processor farm. Have you thought about distributed computing on other people's unused cycles?

I don't know exactly how you do your rendering but I once wrote a three tier system which distributed POV render files via the net. Patrons would run a small program that would retrieve images to be rendered from a server and upload finished images and get a new file. Several of my friends helped me to render some rather long animations overnight and this was like 6 years ago when CPU power was considerably less.

I understand there are several distributed POV rendering program nowadays although I am completely oblivious about them. Mine was a pretty short, efficient collection of programs which only suffered when my web service provider would reset the security settings. What a pain that was.

Good luck with your continued endeavors, I have been quite enjoying the process you are sharing with us.

Tom said...

Hi,

I am glad you are enjoying my projects. I am enjoying bringing them to you.

I have thought of distributed computing. There is NanoHive (http://www.nanohive-1.org/atHome/) which is part of the same company that is developing NanoEngineer 1. And there is Nano@home (http://www.nanoathome.org/) which doesn't appear to be as developed just yet. But there is just something about having my own mini-supercomputer that I can't get over.

To make these animations I run a nano-dynamics simulation in NanoEngineer and save the resulting frames as pov files. I then use POV-Ray 3.6 to convert them into bmp files, and I animate with Movies13.

It is the running of the nano-dynamics simulation that could really benefit from distributed computing. I need to look into it more.

Please stick around. I think things will only get more interesting.

Tom