InstantCAD - MIT - SIGGRAPH 2017



InstantCAD from MIT, to be presented at SIGGRAPH 2017.

So many past and very recent threads discussing improving/speeding modeling. Add this to the pile:

I still contend for version 7, make it the - Optimized for Modeling Update

Meaning - a all hands on desk laser focus solely on the biggest core modeling improvements in Rhino history, within available means, resources, and talent. Still think users will go ape over the idea, if delivered.

One man’s option. I don’t run a long term successful CAD software company and could be wrong…

(David Cockey) #2

It appears that InstantCAD is not directly applicable to Rhino since it uses parametric CAD

Based on the description in MIT News it sounds like InstantCAD is about optimizing a parametric model using analysis software, not about modeling. Its based on a new algorithm for interpolating between points in parameter space based on precomputed data which is claimed to allow real time feedback of analysis results as the parameters are changed.

“It’s too data-intensive to compute every single point, so we have to come up with a way to predict any point in this space from just a small number of sampled data points,” says Schulz. “This is called ‘interpolation,’ and our key technical contribution is a new algorithm we developed to take these samples and estimate points in the space.”

“With this precomputed data, you can instantly improve and optimize the design in two ways. With “interactive exploration,” a user interface provides real-time feedback on how design changes will affect performance, like how the shape of a plane wing impacts air pressure distribution. With “automatic optimization,” you simply tell the system to give you a design with specific characteristics, like a drone that’s as lightweight as possible while still being able to carry the maximum amount of weight.”


Perhaps. Regardless, was speaking conceptually rather than direct application. If InstantCAD represents one theoretical academic approach to specific advancement, Rhino users continue to demand advancements too, which are within means to deliver. Means is the key…

While the underlining technologies (algorithms) may be vastly different, is there not a Grasshopper similarity in terms of sliders and adjust-ability? Very broadly speaking.

Keep Rhino modeling moving forward for V7 was the intent. Obviously other will too in their product.

(Gustavo Fontana) #4

The deformations look no different that changing a basic parameter in a parametric modeler. It looks like they are pre-computing analysis (at a gross level) results of the entire range of parameters.

It you look at the level of interpolations of form that you can get in polygon-based programs with morph targets (also knows as blend shapes), this seems a much more limited subset of that.

I can’t wait to have curve-based subDs where the curves can define SubD patches, and the parameters/dimensions of the curves can be defined/manipulated with Grasshopper.

(Wim Dekeyser) #5

Of course you can wait… :wink:

(Gustavo Fontana) #6

Rrue true. Every new cool feature takes McNeel about 6-8 years to implement half-ass, another 6-8 to get it right. And they are the best in the industry, so I’m not complaining too loudly.


sign me up for that one…


Anything that makes the process faster, flexible, more iterative, smoother, easier to edit, more accurate (thereis a can of worms) etc., etc., get a thumbs up in my book. Everyone just wants a little love, eh.

Still waiting on that - make it all right button…:grinning: