Hacker News Reader

Comment by bhouston | Hacker News Reader

There are too many mature renderers and the competition is intense.

The current popular actively developed ones are: V-Ray (which we use in http://Clara.io), Arnold, Maxwell, Pixar's RenderMan, and KeyShot.

Less popular but actively developed ones are: RedShift, Furry Ball, 3Delight, NVIDIA iRay, Octane/Bridge...

And then the ones that are integrated into the 3D packages themselves like Blender's Cycles, Modo's renderer, Cinema 4D's renderer, Houdini's Mantra, Mental Images (included in most Autodesk products.)

Then the smaller opensource ones: Sunflow, Lux, Corona, Mitsuba, Pixie...

That is a lot of renderers and I am sure that I am missing quite a few.

View on HN

Just as a sidenote: Blender cycles is now becoming a separate standalone renderer itself. src http://www.blendernation.com/2014/01/27/cycles-standalone-re...

Also its being integrated into other 3d packages like Rhino. src http://www.blendernation.com/wp-content/uploads/2014/05/rhin... and http://www.blendernation.com/2014/05/12/developer-meeting-no...

One of my favorites: Indigo

http://www.indigorenderer.com

Nice to see someone knows and likes us :)

I've been a big fan since the days when it was free! Since I do CG as a hobby, can't afford the full version just yet. But, someday, I'd love to!

Do you remember screamline rendering from Intel? Way before its time.

No way, I didn't even know that existed. Just reading up on it now: http://www.creativeplanetnetwork.com/dcp/news/intel-screamli... How did you know about it? I don't think many did.

A similar current company is http://ZyncRender.com, although even today with great bandwidth, cheap cloud computing, and an amazing sync system, they still found it a hard market to crack.

Because age.

While I am programmer, my first love was animation. I had done some cluster computing with thought experiments around huge render farms. When screamline came online a bunch of years later it caught my attention immensely. What a perfect combination of the both!

Aside: I tried doing EC2 type stuff in 1999-2001, it was really hard, too much resistance for running anything "sensitive" in the cloud.

What a great way for Intel to sidestep the artificial scarcity and release cycles to the world instead of chips! They could prototype new hardware, use "crippled" cpus that they couldn't sell, I thought it was a great vertical integration play that sidestepped many of the ridiculous economic factors at the time. Eventually it will get there when we have massive bandwidth, AMD and Intel if they are still around will sell mostly cycles.

I was really bummed to see screamline go out of business. It was like a sad canary.

Intel Screamline Rendering Services:

* 2001: https://web.archive.org/web/20010722053709/http://www.scream...

  Try Screamline Rendering Services and your first 500 
  frames are on us!

* 2002: https://web.archive.org/web/20020124110135/http://www.scream...

  Screamline Rendering Services, provided by Internet 
  Computing Services, has enjoyed working with our 
  customers in the computer graphics animation industry to 
  provide outsourced rendering services. 
  
  Due to a deteriorated funding climate for Internet-based 
  businesses and unfavorable market conditions, Internet 
  Computing Services, part of the Intel New Business Group, 
  will exit the outsourced rendering services business, 
  effective October 31, 2001.

Currently, Intel has the Embree ray tracing project, which is doing lots of interesting things with ray packet algorithms and SIMD compilers.

http://embree.github.io/

V-Ray adopted the Embree project into their renderer with V-Ray 3.0. :)

As have Pixar for PRMan 19 for some of the BVH build stuff :)

Adding Guerilla Render to your list in the animation space... http://www.guerillarender.com

How many of those take advantage of GPU floating point power? This is no longer a nice to have sort of thing, but a critical one, since as the new Mac Pro shows, there's going to be orders of magnitude more GPU than CPU power available in newer workstations.

I've seem some, like Furry Ball, that support only CUDA, but very few are vendor agnostic with OpenCL. I hope AMD tries to fix this.

I don't think it's a critical feature for a commercial renderer.

- CPU renders typically might use 32Gb memory (or more).

- most rendering is done on a render farm which won't have GPUs (unless you maintain a gnu render farm).

There are good reasons why commercial renderers are avoiding the GPU, at least for now.

Xeon Phi may well cement this further. Knight's Landing, due in 2015, is going to feature 72 Atom cores (288 threads) with AVX, socketed in a standard Xeon motherboard.

In situations where your problem domain doesn't fit comfortably within the memory restrictions of a GPU, or porting a legacy code base difficult, this could be a very interesting option.

http://en.wikipedia.org/wiki/Xeon_Phi#Knights_Landing

It is a non-starter for the successful commercial renderers because the majority of their market already has render farms with standard CPUs. Also studios generally use a mix of tools and if only one of them is optimized for Xeon Phi's that isn't enough of a motivation to spend the money on Xeon Phi's.

It is generally a no go in the mainstream rendering market, although cloud-based renderers can use specialized solutions.

I think that's the way as well

You may be able to accelerate some steps in a GPU, but rendering usually involves "a lot of data", if you add all resources for a given scene it might be on the order of gigabytes

This is not like bitcoin mining where you have to do a lot of math in a small batch of data.

Cycles uses GPU and CPU both for rendering.src http://www.blenderguru.com/4-easy-ways-to-speed-up-cycles/

Octane is GPU exclusively and uses CUDA AFAIK.

I've used Octane, which is a GPU-only renderer. Unfortunately, it is VRAM-per-GPU bound. In other words, you cannot render any scene larger than the usable VRAM per GPU. High end video cards with lots of VRAM also tend to have more than one GPU, so the VRAM-per-GPU figure is actually (total VRAM)/(total GPUs), with slightly less than that actually available for the scene.

I've also experimented a bit with Lux which has a hybrid CPU/GPU mode. However I've found it isn't necessarily any faster than CPU only on my system (which has a lot of CPU cores) and it isn't as stable.

AFAIK, there are no video cards currently available with more than 6GB per GPU, since something like a nVidia Titan Z with 12GB has to share that between 2 GPUs.

It's conceivable that as GPU rendering becomes more commonplace we'll start to see manufacturers loading more and more RAM only high end cards, possibly at the expense of compute units if power consumption is a problem. After all, a render farm with many separate cards is just as fast as one card with more compute units, but VRAM per GPU is currently a hard limit that will affect anyone rendering very large, complex scenes.

For doing production VFX shots, unless there's fur/hair or volumetrics involved, you're pretty much IO constrained in terms of pulling in textures and paging them in memory.

Often scenes can have > 300GB of textures total. Out-of-core rendering is not something that GPU renderers can do very well yet (redshift can do it a bit), so the 12GB limit on GPU memory is a huge problem. Arnold and PRMan both have deferred shading which means for GI they can re-order texture requests for similar texture mipmap levels to reduce the swapping of textures needed. I'm not aware of any GPU renderers that can do that as of yet.

On top of this, geometry assets are always overbuilt these days, with millions of polys/subds, and often multiple displacement layers on top of this. This geometry requires even more memory.

And the biggest studios are rendering deep output, requiring even more memory for the output files as more of the render samples need to be kept in memory for the render duration.

And while GPUs are faster, they're not THAT much faster for raytracing - a Dual Xeon top end CPU can almost (85% of ) match a K6000 at pure ray/triangle intersection speed. Where GPUs can win is the fact it's a lot easier to cram multiple GPUs into a single workstation than get a 4 socket CPU system.

But in a renderfarm, this isn't an issue.

In a traditional render farm that's not an issue. As the new supercomputing centres have shown, there's a huge push to add massive amounts of GPU power to these because it's more cost effective.

I'm not saying GPU rendering is easy, it's way harder to do correctly because of memory constraints, but the performance of GPUs over CPUs (compute per watt, compute total) is a gap that's only getting wider. It's easy to add a few thousand more shaders, just add another card. Not easy to strap on another 12-core chip, you need to engineer from the ground up for that.

A GPU also has access to all system memory via DMA, so that shouldn't be an issue. It's just going to be hard to coordinate that data transfer in a performant way.

In a traditional render farm that is an issue, as the vast majority of renderfarms for VFX companies are CPU only. renderfarms aren't just using for renderer rendering, they're used for comp rendering, fluid sims (often using up to 96 GB of RAM), physics sims, etc.

Very little of this is GPU-based currently.

Where GPUs are starting to be used is on the artist workstation to speed up iteration/preview time. But not on renderfarms.

GPU memory access across the PCIE bus is ridiculously slow for smaller jobs. It's often the case that the slower CPU can do the work before the cumulative time of copy data to GPU, get GPU to do work, then copy it back again has been done. For longer jobs of smaller data, GPUs make sense.

GPUs also have much higher power draw than CPUs and produce much more heat, which is a huge issue for renderfarms with close to 100% utilisation at crunch time.

There's also http://appleseedhq.net

There is also the Renderman compatible Aqsis, although the last time I looked, development on that had slowed considerably as they work on version 2.0. I got the mailing from Pixar about the pricing changes and new version yesterday. I'm curious about the technical changes in Renderman and the new RIS architecture. Does it still have RIB? And SL? I can't imagine those going away. I'm also interested in the non-commercial free version. One can never have too many renderers to play with.

Another integrated renderer is the one is Lightwave. I believe that Blue Sky and Dreamworks still use their own proprietary renderers. Sony is using an Arnold based renderer.

Time to start the race to the bottom in renderers market!

Anyone know the status on POV-Ray compared to other open source renderers? I jused to play with that a lot.

Reply on news.ycombinator.com