How much power to supercharge rhino


I am an architect. I spend my work time modeling and rendering buildings. Houses/Apartments and the like.
I was taught rhino in school for the past 4 years, but have really spent the past year attempting to understand the correct way to use rhino.

I sought rhino streamlined methods, meshing, small saving, keeping polygons low, ALAS!

The time has come where I still find my file sizes ballooning over a gig…

I am at a point where I realize all my effort in attempting to " LOWER FILE SIZES," all the innumerable, countless, timeless zones of limbo I’ve spent tinkering with the model, to diluting the display options down to the bare skeletal minimum. Dealing with the matrix of wireframes over a fully loaded model. can be avoided if rhino would just have some more torque behind it to pull the weight of everything I’m asking from it.

I feel as though I am in 1995, with windows 95 and an AOL trail disc, trying to download a movie.
IS IT POSSIBLE? YES. can it be better. YES.

Now. I am asking this. HOW MUCH RAM, how much brute force worth of xeon processors and how many top notch video cards do i need to render farm into this rhino software, to get my functionality up to 10 gigs.
If i had up to 10 gigs worth of modeling capacity. Regardless of polygon sizes, regardless of sel dups, or all the constant constant constant constant termites of rhino’s workflow, I could just flatten them all out and ignore them through sheer processing capacity.

I don’t want my screen to lag, I don’t want to work in wire frame, I want one monitor to be breezing by in fully rendered view, one monitor to run a full blown 2,400 by 1,600 resolution Maxwell fire, and i need to be able to model with no hiccups whatsoever.

What. will this take. Thank you. thank you. thank you.


You probably can’t.
Rhino was designed to work on humble and moderate hardware. It was not designed to take advantage of “supercharged” hardware.
Good performance becomes more about working intelligently than by just slamming along and pushing harder.

You’re in the right place. There are some very talented and productive Rhino users here. Hardware is part of the solution but it’s not the biggest part.

Ask specific questions about specific problems. General questions are either ignored or answered with general answers which never directly translate into solving specific problems.


hahahahahaha, thats fantastic man,
ahhhh that’s precisely! PRECISELY- i’m glad i came across that clear,
BUT SERIOUSLY, lmfao, oh man, ^5 man, you got it. : )

John. I asked my specific question earlier.
It was my first question, and I got abundant directions offered to me,
Different programs, different software, and I have spent the past few weeks checking out youtube videos,
Downloading different programs, from revit to others,

I start creating a basic block. in each program, and i copy paste it, just to see how much it can handle.
I copy paste, until i have 10, 100, 1000, 100,000 of the same block.

nearly every program, with the exception of revit, which barely maintains after a few hundred; is capable of sustaining smooth usage once 10,000 of those units are in place.

When you tell me I probably can’t, Its a sign for me that I am at the limit.
How then do pixar and dreamworks do what they do? Do They also suffer with laggy computers?

I read that in the new zootopia, each animal has more hairs than every character in frozen combined.
I read before that- that elsa’s ( the blonde rapunzul princess girl ) ( she’s impossible to miss )
had more hair strands than all of monsters inc characters.

My point is, that they have this consistent benchmarks they keep reaching with every evolution.
I understand completely there is " working intelligently " but time is also an aspect here.

Sometimes I need things quick. Sometimes a client wants a new couch, table, desk, chair, entire backyard redesign.
When I drag and drop things from 3d warehouse, I don’t always have the luxury to drop it into a blank file, and dissect it, and fix it, and take 20 minutes doing all of this clean up work.

I need to drag and drop. done.

I have read about mayas new nvidia linked rendering software; lray, its maxwell all over again. Blender has the same attributes. It seems like the industry is just re-hashing the same stuff.

I don’t design watches, or bottles, or cars. I design houses. Houses full of furniture. Apartments with multiple rooms with furniture- With an accumulation of thousands of things.
I know there are blocks, and proxies for other programs, But why do i have to work like this?

In photoshop, no one advises you to place a proxy and pretend its red because photoshop can’t handle all the colors at once- you don’t have to create half of your sketches in one file, and another half in another file, and then at the end you bring them together. no. You work all at one place, and its smooth and cohesive, and it never slows down DURING the process.

It takes its time once blurs or heavier options are chosen, sure. But in general- its quick and snappy.

As far as I can see. There are no programs on the market that can effectively, efficiently cater to a quick rendering workflow. Whoever has understood how to navigate these different programs to create a quick rendering workflow, has most likely done so through personal libraries of extremely low poly objects. But even then those libraries are finite and closed off.


I think that there’s no level of hardware that will compensate for careless design when doing large projects.

As far as a 10 Gig file, that file could be saved 2 different ways; the more compact of which will take just as much memory as the other once it opens.

Of course, things such has having enough RAM as not to let the computer use virtual memory would help more than a faster processor. A server motherboard with a lot of RAM slots and a graphics card PCI express slot would help let you keep more in memory. Once you run out of physical RAM, it will be tough going. Though, many gamer enthusiast -grade motherboards can take 32 GB of RAM.

There is also functional limit to how many triangles can even be rendered by a graphic card, so things like simplifying rendering meshes, and hiding unseen objects helps. Rhino can simplify objects to maintain a frame-rate; it helps.

Lowering your maximum undo’s would help too.

Rhino V5 is still lightly threaded, so you want a fast clock speed and a CPU with a large cache more than more cores–until you render something.

My friend worked doing a subsystem on a remodel for a large airport; he asked me if there anything I could think of to keep Revit and Autocad from crashing, so it would seem that other programs have functional limits as well, and perhaps worse ones.

no level of hardware
lightly threaded
functional limits


I have 64 mb of ram, with an i7 3930, and a gtx 980 ti, which truthfully are the best i could buy from fries. Its a custom build.

I was thinking maybe i had to upgrade to a professional grade computer, with the expensive xeon processors and sli 8 graphics cards together to …madmax through the bumps. lol

in any case miss brenda, john, everyone. I do appreciate everyones efforts.

That pretty much sounds like the best set up you could have gotten already. So I think there’s little to be gained on the hardware front.

Rhino6 has a number of display improvements (and more are being added still), so you could have a look to see if the framerate is any better. Did you join the Rhino WIP (ie. Serengeti) program?

I just saw you knew about this to some extent but give it a try, it’s quite easy to do.

I’ve worked a lot with train interiors and I know your frustration, I had 3-4 gb files. What I did was to use VRAY and blocks to get a smooth experience. Maybe this wont work in your case but I think it will to some extent.

What i did was that I looked at the model and identified objects that are used a lot, in a train this is pretty easy, most of the interior repeats every 2-4 meters but I think that also a house has a lot of repetetive objects like windows etc.

Then I used VRays fantastic function Proxy rendering. It’s a function to ease the load on the CPU (or more the RAM) when rendering so objects are loaded into the rendering as they are needed. Another vital part in this function is that Vray replaces the object inside Rhino with a low-poly version of it (you can decide how many polygons you want for each indivdual object).

So what i did whas to make rhinofiles with all parts that repeated, made a layer with the real version of the object and a layer with the Proxy version. Then I rebuild the scene with the files, imported as blocks.

The result is a rhinoscene with the option to make parts lowpoly or highpoly. Then of course when you render it all goes super fast. A bonus is that since your scene is built from blocks where the blocks are repeated you will get much smaller filesizes.


Softwares for VFX build meshes in a different way than Rhino so you will have less polygons. Also all the effects with hair and such never has a pretty preview, it looks like shit before you render and rendering takes huge amounts of time on huge farms.

1 Like

I’m sorry, but I really have to interject here. I really do not see how your file can be 10gb, there must be some serious optimisations that could be made to it.
The problem here is likely your modelling approach, not the CPU/GPU limits of the hardware…

1 Like

10 gig it’s a big one file, use external block to lower your file. Rhino is the most powerfull 3D CAO modeler. i have HP workstation and the screen is limit by the polygon count.
if you have juste simple object like wall it will be good but if you use NURBS curvature the NURBS are redraw all the time to have the perfect representation.
a 10Gig model in architecture is create by contractor not architect.

Sli must be the solution but i think 3 graphic card Geforce a big xeon and 64go RAM, but not sure

First of all my friend, don’t apologize. Your opinion is entirely welcome.
Second. My file is not 10 gigs. I want free functionality up to 10 gigs.

I find everyone here is content with what is available, if not content then…lets say " realistic "
for what is available.

Personally I am exhausting abundant work hours simply deflecting and navigating avenues to get the program to work.
I think about 60 percent of my work, if not more, is just dealing with rhino.

and optimisations? SURE- But I doubt anyone is hearing me clearly.
Let’s say I am the world’s worst modeler. Let’s say I have no concept of blocks, or anything.

Maybe I just want to copy paste the same thing over and over again, until i hit the millions mark

Maybe I want to scroll through unbelievable amounts of walls i’ve extruded for no reason.
without rhino slowing down.

Is this even a possibility?

I get that its almost an abstract question, I recognize the measures we take to lessen the strain on the software, but my entire objective is different. Instead of lessening the weights, Is there a way we can add some horsepower to rhino. Thats my entire idea. The closest anyone’s come to has been Jjin. And Jjin if your reading this I’m definitely gunna check it out.

anyway. to summarize.

YES. theres lots of ways to lighten the load, and work around the capabilities of rhino.
I feel like i’m repeating myself here, but

Instead of throwing people overboard from our little canoe of rhino to go faster, can we turn this canoe into a cruise liner? you guys get what i mean? Can I just make Rhino Lift more?

Greetings Mun,

We understand the question: “How can I get Rhino to do the work of a god?” And I do sympathize - to a minimal extent.

More - Faster - Better - Stronger - Oh the glory!

The software companies are just repackaging and reselling rehash - yep - consumerism rules!

What you seek may be a visionary application of big-data technology with Rhino as a small viewport on a virtual world.
Rhino meets SecondLife or some such thing.

I bet the motion picture industry knows how to do it - You have a venture capital firm in your back pocket?

1 Like

I wished Rhino could handle files upto 1GB and keep the frame rate at a high level. I’m working on large train interiors/exteriors often and frame rates around 0.5fps are horrible. The workflow is nerve killing. And last I tested v6 it was not better. Realistic ideas for a faster display are sayed (mesh only based), but … .

You are asking a valid question, but proposing the wrong solution. You need only enough RAM to load your scene, plus necessary scratch/swap space: a good rule of thumb is 2x what your biggest model requires on top of your other software, etc, but you don’t necessarily need to go that far. This will prevent huge penalties when you try to do work with full memory. But beyond preventing memory chokes, more RAM won’t make your machine faster for interacting with a large model. Neither will many-core CPUs, since most CAD operations are single threaded. If you want smooth frame rates from very large files, you need a bigger faster GPU, with more dedicated graphics memory. Nvidia’s Quadro M6000 is the biggest, most powerful pro graphics card available, with around 2,800 GPU cores, and 24 GB of very fast frame buffer memory optimized for OpenGL (2x faster than GeForce) and CUDA ray tracing. It’s an expensive piece of hardware, but it will tranform your workflow. So, the question really, is not “How can you make your system faster?” but, “How much do you value your time?”

1 Like

And get 64gb of fast RAM and make a ramdisk

I value my time, very, very much.
I checked out the graphcs card. Man you don’t know how thankful I am for this.
I have a question for you.
If i got TWO of these cards, would it help? Even further still, if i got FOUR. would it help? Could I sli four of these?
Can I have maxwell fire dedicate itself to one graphics card? and another just to work on?
Thank you. Thank you. Thank you.

Finaly- Prehabitat… I have 64 gigs of ram… what is ramdisk?.. I feel like i should know this already…

Ramdisk uses your RAM as a harddrive so you get a super charged harddrive that forgets everything when you shut it off. Can be good for some software, not sure on how much it affects Rhino though.

Yes you can use multiple Quadro M6000 in SLI mode, however, you will need a lot of power for that computer, also a very large chassi with a lot of cooling/fans. Also note that cards in SLI don’t give 200% more like 140% (edit: it’s more like 150-180%) if i remember correctly. (yes the M6000 can be SLI’d) (yes Rhino can use SLI) (breakdown on increase in performance with 2-3-4 way SLI)

If you want fast rendering you should have a look at Octane, especially if you go for that many GPU’s, you will have quite swift rendering directly in the viewport.

Added links.

Unfortunately, John’s reply above is probably fairly appropriate. Still, there ARE some things you can do to REALLY get close to your stated goals.

RHINO is one of the MOST capable tools on the market. If you SUCCESSFULLY achieve your intents using RHINO, you’ll have achieved a system that will perform capably, as you bring other tools into your arsenal. I also use AutoCAD (YES - ACAD ! ), Inventor, Cinema-4D and 3DS-Max.

1.Processor- YES multi-core XEON processors is appropriate, but let’s pause for a moment and touch on MAIN Processor cores. This way you can know how they best fit in, and THEN where to go from there. First, let’s remember that most CAD Software systems are still single-threaded. There may be some exceptions, but those will be limited to specific implementations so really don’t affect our conversation very much here.

Being a single-threaded system, RHINO will make effective use of ONE core, with HYPER-THREADING (Intel-Core processors) largely a non-factor, claims to the opposite notwithstanding. SO you’ve noticed what “seems” to be a bottleneck. Yes, you’re right. SO- get as FAST a multi-core XEON as you possibly can. Overclocked STABLE systems from such manufacturers as those offered by BOXX technologies,@xi and the i-x2 workstations from Mediaworkstations are great examples. They even offer MULTI-core options in both the XEON and intel-Core worlds that are truly astounding ! Lots of cores available, running REALLY fast. Current upper limits run around 3.4-4.0 Ghz, with some base speeds as much as 4.0 and with overclocking, if you’re careful to go with a fully tested and certified unit, follow cooling and related recommendations, your chosen vendor can help you optimize a result as close to your desired goals as can be currently achieved.

Ok, so RHINO is primarily single-threaded. Multi cores ? Yes, but for other reasons I’ll get to shortly.

2.RAM- Get AS MUCH RAM as you can achieve, based on the limitations of your chosen platform. High speed RAM - LOTS of CERTIFIED DDR-4 RAM at the highest bus speeds your platform (as recommended by your chosen vendor) will handle is the ticket. Remember, though, as you work with your vendor - let them know your application (not software) and discuss reasonable ways you plan to grow your toolkit. RAM above certain levels may be of little value. We see 32Gb typical in some standalone workstations, so that’s a reasonable starting point, but if you plan on doing more than basic 3D CAD, you can reasonably consider more, depending on additional services you’d be running at the same time (like rendering and animating, etc.) . More on that when I get to VIDEO. Work with your vendor. Their recommendations will be very important.
3. PRIMARY and WORKING storage- With available capacities, PCIE-SSD is the only viable option. Yes there are hi-speed physical drives out there and some REALLY COOL RAID options available, but you identified a system-lag problem, and the HD bottleneck is actually a probable culprit. LOTS of RAM will reduce file swapping but a hi-capacity PCIE-SSD is the answer.
4. Video- Work with your vendor on this, but you’ll find your most value in a video solution for RHINO that optimizes open-gl implementation. NVIDEA and AMD offer worthy solutions. While you’ll probably want to avoid a “Gaming” card, I’ve found RHINO’s use of NVIDEA’s implementation in their gaming cards to be a very viable smaller-scale approach. HOWEVER - and this is key - you’ll want to focus on one of those 2 (NVIDEA or AMD) in a Workstation class video implementation. Even so, however, more than 2 video cards in tandem (whatever your chosen protocol) is probably as much as you’ll get any value out of.

With that said. Boxx industries offers a render-station that piggy-backs onto your primary workstations, with PLENTY of capacity to expand beyond the 2 you have in your primary workstation. This will ACHIEVE your intent of having top-resolution running in real-time on one monitor while you’re designing on another. Real-time render solutions really shine, and RHINO’s implementation is among the BEST I’ve seen. While that uses multi (like “CUDA” ) cores on the GPU (Video card), for FULL rendering, you’ll need MAIN PROCESSOR speed and cores. You don’t need the external render-box to do this, but its a really sweet option when you’re looking at hi-end time and sunlight renderings, and the various animations you’ll be doing for concept presentation. ONE hi-dollar contract will make your purchase well worth it.

This is where the multi-core main processor can really shine. Hi-end renderers mainly use the “other” available cores on your main processor, so some processors with 12, 16, and 24+ are not unreasonable - with this excerpted from an on-line ad:
44 Core Xeon™ Workstation‎
Dual Xeon™ E5-V4 CPUs - 8-44 Cores,‎ Up to 512GB, Low Price High Quality‎‎
If the cores are in your primary unit, that works, if they’re in your render-box, that works, and if you opt to network unused-available cores in other units in your office that works, too. I’ve successfully avoided “cloud” services, as ALL of my clients’ products are very proprietary, and NO, promises of "secure site … " really don’t impress me.

So- without making any “specific” recommendations ( read "endorsements :wink: ) I think you can see that with a little pointing in the right direction, you CAN get there from here !

Final point - while you’re researching all of this, be sure to read everything you can that’s offered by the folks here at RHINO. They have some good starting points well worth considering. I’ve not found a SINGLE instance where a recommendation for my RHINO faulted at my other tools (Inventor … etc.). Their Technical support team is among the best I’ve dealt with in my 35 years as a Product and Equipment Designer for the Offshore Oil and Gas industry. I’m a 10-plus year RHINO user in Mechanical Design, Modeling and Animation, and find it to be a VALUABLE tool in my arsenal !

Post back with results ?

Thanks - and good luck !