I use Rhino for visualization work, for example train and air plane interiors. The 3D data can be quite complex, most if I get the data from the engineers.
Well, McNeel just needs to fix this problem, or at least fix it partially. thatâs frustrating.
Hi John, I know this is a tough subject. And all I want to do is to shed some light on the professional users point of view. So please read on:
Complexity as of 2010+
Rhino is for designers and architects ++ which means working with engineers on complex projects, and all projects are complex once they move towards production, so that means handling a lot of data, and that requires a lot of horsepower. Today lots of components are built in 3D with lots of details, working with a small boat can mean having two large 3D modelled engines, and or lots of small components around the steering wheel, that alone are much more complex than the rest of the boat.
A designer often focus on simplicity, so the curves and the surfaces can be very light, but donât confuse that with no need to handle complex parts. Rendered, glossy, with shadows and reflections at smooth and slow mesh settings. Evaluation of the whole is key.
If things are slow then the users have to spend a lot of time hiding and unhiding things, and organizing what is important to see at any given time is a time consuming task.
This file is not a very complex model with clean, smooth surfaces, no bad objects etc. And in shaded mode it gives me 14 fps. But in Ghosted I get as little as 4 fps. And this is on an i7 system with a Quadro 4000. Needless to say, 4 fps is not fast enough to act as realtime.
Multiple pipelines
I just used the directx-opengl example to illustrate that Rhino handles multiple pipelines. I did not say that you should make a DirectX pipeline too.
Light, shadows and reflections
When Working with architecture light is key. Being able to navigate the scene when the sun and skylight is on can be crucial for the ability to evaluate the design. In example, this is an highly optimized scene for light evaluation, and it consists of only 134 meshes with a total of 1.154.718 polygons. In what I call âRenderedSpeed modeâ with no complex lighting, nor shadows, I get 123 fps. In Rendered mode with shadows and only the sun I get 14 fps⌠(that is too slow for silky smooth evaluation), and with sun and sky it drops down to 4fps.
Needless to say, you need to find a way to speed up those calculations, a way to cash them to memory and reuse them, as the sun shadow map is static unless the sun is moved, and so is the skylight data, or at least find a way to not show the skylight shadows while manipulating the view. Learn from Neon and do stuff in post.
Blocks as fast as pointclouds
To me the key question is: How can you make a block or a group of objects be almost as fast to handle as a pointcloud or a complex mesh? (Compare 10K single points to a pointcloud, or 1k simple meshes to 100 complex meshes and youâll understand what I mean)
A block is just as static as a joined object, a pointcloud or a complex mesh, and should be considered as a single object. Iâd say that 90% of the users who work with blocks would rather have them display very fast, than have them âopen for editâ very fast.
Next gen monitors
Dual 4k monitor setups are here now, and will be common during the next few years, so start optimizing Rhino 6 for this. Inexpensive laptops are also coming with UHD displays, and that is a lot of pixels to draw, so Rhino will need as much speed as possible.
The slow systems would benefit the most!
Everybody would benefit from speedups, and mostly those with little horsepower! In example: Getting a 3x speed boost If you only get 5 fps is a HUGE step, but getting 3x if you already get 30fps is barely noticeable.
So it is in line with your philosophy, not in contradiction to it.
2 Likes
Thanks for taking the time and effort to explain your situation. This sort of discussion is exactly what @bobmcneel and the developers need to hear, and what we hoped would happen with this forum. It helps us decide what needs to be worked on and what direction to guide the development priorities. Donât get me wrong, I canât provide any assurances that this will be the direction we go, but it makes a good argument for input for those working on the big picture.
Very good. Thanks again
1 Like
If youâre building a top-end system for large files like that, Iâd get an SSD for the OS and current working files.
Yes the SSDâs properly configured with the other hardware are AWESOME! Also the way they work I donât think software matters, it just moves the data 10x faster.
As one of your customers I certainly would not regard myself an expert in this subject. Speaking idealistically not realistically I think software should only cost about $35. But I understand thatâs not realistic today.
the lower price of your product combined with apparent quality and versatility of the product and support of it caused me to buy it. Keeping your product at about the same place it is currently price wise, which is way way way way way cheaper than Autodesk is important, if it was a bit more it wouldnât bother me. And keeping your customers rather than losing them is more important
So for financial reason, say you needed to create an add on called say hypothetically âRhinoceros Turboâ or something like that. I would be willing to pay around $300.00-$400.00 for that special upgrade package that would basically allow the software to take full advantage of higher end hardware. and this software package would awesomely help retain your customer base, who will likely leave you as they develop in to more mature and experiences modelers, developing more complex and demanding models for various fields.
I cannot see staying with Rhinoceros software if they donât fix or at least partially fix this problem.
Rhino could only cost 35$ it you broke it down to 35$ components(plugins) so each time you would need a new tool you would have to pay for it. That could work, in theory.
Keep in mind that Rhino is not a toy. It is a tool. And not only a tool, but more like a really, really big toolbox. A factory even. Where you can design, adjust, build and manufacture your ideas.
It is for professional people, who charge more than 35$ an hour. None of those would consider a 35$ tool a serious tool, because they would know that the price would never ever cover the development time required to make it stable. And stability and knowing that the software will be around for a long time, is crucial.
So I presume you are really wishing for a slimmed down, lowcost version, or a personal learning edition.
Would that cover your need?
Hi Justin,
I find it hard to believe that after investing several years in to the use of Rhino that users will leave by their own choice.
From a recent conference in London it became apparent that Rhino is the first tool in the process. It is not the Swiss Army knife of knives. There are few other choices in the early conceptual surface modelling phase of a large project that offers the same as Rhino in terms of ease of use and speed of turn around. Those companies who had the resources might develop their own tools to then manage the rest of the project from within Rhino and those that didnât just used other third party plug-ins or proprietorial software.
As it has been previously mentioned Rhino is used in so many different disciplines its hard to find the right balance of what features are provided by McNeel as standard and at what cost and those that are specialized in by Third Party Plug-ins.
Incidentally I use two graphics cards to drive two monitors. My system is 7 years old now and still man enough to handle large complex models. I found the bang for buck performance of driving one monitor with one card twice more preferable than getting a massive graphics card to be halved in resources by driving two monitors. There is no SLI bridge
No thats not what im talking about. It doesnât matter. im being purely philosophical and idealistic, not realistic. it doesnât matter.
Just ignore my statements about $35.00 dollar software. its very very philosophical and makes absolutely no sense at all to anyone, and should not be attempted by anyone, at all, unless you know exactly what your doing.
Since Iâm sure none of you here know how to make high quality software for 35 dollars, you should not attempt to do so.
Hehe, and if it was easy and if everybody did, then the world would have been full of it!
Being philosophical is all great and very important, and so is defining the value and meaning of 35$.
For it to work drivers and OSâs and hardware and all had to simply just work all the time, with out updates or upgrades or other time consuming things, or for the application to run on a platform developed by some bigger company. And then it would have to be marketed with out effort, and be distributed with out any distributors.
I do see the need for a simple sketch tool for the iPad that could have been a strongly limited Rhino for Mac tool.
A tool to sketch while on the go. With basic 2D and surface tools. That would make sense to me.
Btw, take a look at www.moi3d.com that might be just what you need.
The software is a sort of swiss army knife to me, and I look forward to CAD software and software in general sort gelling together in the future into a big swiss army knife kind of a thing. Similar to the way phones and tvâs and radios and computers have gelled together into one blob of goodness. That what âversatilityâ means. Rhino markets itself as a swiss army knife of CAD packages, thats what makes it cool.
I envision new super efficient business models developing around the planning and developmental power of modern computers using swiss army knife like CAD tool packages, and these companies will redefine our ideas about wealth and custom of living.
The only reason these companies are using the software the way they are is because thats just where the technology is today, if the tech was better they would be doing thing differently. And also learning curve is pretty steep.
I am basically concerned that some bottlenecks within the soft-wares own architecture will prevent me from âfleshingâ out my designs, and studying my parts, and flow and movement of them.
Addressing these weaknesses in the software by building a âTurboâ plugin that fully uses the hardware would correct this and make McNeel a real competitor with Autodesk.
Time investment is meaningless to me, all the CAD software is about the same, its just math. If the software wonât use the hardware then I may have no choice but to find another CAD package. Having said that, I am certainly stuck with Rhino for now. and I am just working on a Thinkpad T400s w/ 4GB RAM in it, which is suitable for learning the software. but certainly will not be suitable for developing complete models with moving parts.
And this is obviously just a money issue. obviously it makes no sense at all not to fully use the hardware. Their are cost involved in addressing these issues, and your just debating whether its worth the investment. I think that it is, why buy two CAD packages when you could buy one. And all i can say, as really more of a student at this point in time, is that I am willing to pay for some kind of âTurboâ add-on that causes Rhino to fully use the hardware. Honestly even at say $500.00 for this add-on you guys are still way cheaper than Autodesk.
Like this:
The only issue is this is 10 times the price of a standard Swiss Army Knife
Looks very nice. I wonder if it works well for bigger complex models.
Yes its hard to know what tools to implement. So far as I know I can already build anything in Rhino and create a solid out of it. from their I think about three things.
- what are my objects made out of
- how will they act and react in environment (wind, water, earth, fire, people, sunlight, electricty, gravity, electromagnitism)
- ergonomical planning is it too big too small do the objects flow well together
The Dim tools could be better, maybe add an âauto Dimâ and an option for the Dimension text and numbers to stay readable at say a 10 or 12 point size rather than behaving like a 3 dimensional object in space so they are completely unreadable. This is why CAD is interesting to me Computers are very good at generating useful numbers. I think Rhino could beef up its number generating abilities.
I see your point, and I understand your point of view, and it is in many ways Googles approach, buy up and give it to the people⌠but it contradicts your fear of the big companies owning too much. If Google bought Rhino what would it be? And do we really want that for our future?
Btw, I saw Google just purchased Boston Dynamics. And I wonder why Google is interested in owning the development of hardware drones.
EDIT: This became REALLY of topic. I trust you know that I am one of those who work hard to get McNeel to make Rhino to run as fast as possible on pro hardware.
Cheers 
Is any of this really happening to you or are you just theorizing here??
No. CPU processing is math. Software is about creating an interface between that math processing power and human beings, which allows said human beings to understand, grasp and manipulate otherwise very abstract concepts. Lots of different approaches, sophistication levels, learning curves and price ranges in a constantly evolving social, economic and technological environment.
âMitch
I canât build detailed files on the computer I am currently on. But using a little multiplication, I take what I am doing and figure I am going to end up with some pretty big files.
This conversation started because I was buying hardware for the software, and after questioning people about this obvious subject, what parts for what kind of files on the software, found Rhino to be extremely dodgy about the subject.
anyway after much persistence I learn Rhino is not built to use the hardware available to it, that its display pipeline is a bottleneck, and that is a ceiling that the developers need to break.
I hate bottlenecks. I hate buying hardware that only works at half its potential because it connected to some old port that wont allow it to work.
anyway, this is all obvious stuff, of course software should work on hardware.
I have another question. Letâs suppose that I have two graphic cards on board in my computer, that are not connected neither in SLI or Crossfire (One is used for display on monitor and another one used for rendering in Blender 3D, for example). Will Rhino recognise that I have two graphic cards installed and will let me choose which one I use for the 3D display ?
Yeah, with all those limitations, bottlenecks and general dodgyness, Rhino really is a piece of junk - itâs a wonder anybody still uses it⌠Crazy, huhâŚ?
âMitch