GPU Render links

- Hou er rekening mee dat in de loop van de tijd de internet pagina's zijn gewijzigd of zelfs gewist. Ook kan de website uit de lucht zijn gehaald. Meetdatum van deze links is september 2015. -


"Rendering is een oefening in het berekenen hoeveel computerkracht u bereid bent te investeren om de realiteit zo goed mogelijk te simuleren. De kosten betreffen zowel de hardware als de software plus de leertijd en de productietijd in Euro's."

bron: fxguide The ARt of Rendering by Mike Seymour, april 10, 2012.


Nvidia graphics

GPU processoren zijn er niet alleen om data op het scherm(en) te tonen, maar ook om grafische afbeeldingen samen te stellen, in de Cloud te werken of zware computer berekeningen uit te voeren. Of om real-time Epidemic forecasting te doen (Lancaster University UK). Of voor cosmologische simulaties met de Titan supercomputer.

In Australië is de eerste Data Arena uitgerust met 27.000 CUDA cores, waarbij u niet alleen kunt kijken, maar ook kunt horen en de data ook 'kunt voelen' door 3D visualisatie technieken in een 360º data visualisatie kamer (Data Arena). Deze is ondergebracht bij de Faculty of Engineering and IT, de software is Open Source.

Datasets in 3D

Sterke schaalvergroting van GPU ten opzichte van CPU toepassing met een Bayeslan posterior density functie voor het voorspellen van ziektes in Engeland.

Vertikaal de kloksnelheid
Horizontaal het aantal processoren
Zwart GPU
Donkergrijs naive CPU
Lichter grijs Vector CPU

Picture shows the results of a scalability study performed using the UK 2001 foot and mouth outbreak dataset and model (Keeling et al. 2001). The figure compares the wall clock time for GPU, non-vectorised CPU, and vectorised CPU implementations of the Bayesian posterior density function calculation. The test system was an HP Z820 workstation with two Intel Xeon E5-2650 processors, and two NVIDIA Tesla C2075 GPUs, though only one GPU was used.

The GPU offers a 380 times speedup of the calculation compared to non-vectorised CPU implementation running with a single CPU core. Moreover, it outperforms by a factor of 2.5 the vectorised CPU version running with all 32 logical CPU cores present on an HP Z820 workstation. The result is that overnight forecasting of another outbreak of foot and mouth disease in the UK is now possible, and can quickly be put into action if the need arises.

The Best GPU Render?

Vergelijking tussen enkele GPU render programma's. Waarbij snel duidelijkheid wordt verkregen over de voors en de tegens van deze korte opsomming van merken.

Expert Advice Tips and Tricks

GPU voor een ander vakgebied

Vijf stappen om met Vegas Pro 11 (inmiddels v. 13) en grafische kaarten om te gaan. De nieuwe GPU mogelijkheden in Vegas Pro 11 kunnen een groot verschil opleveren in de dagelijkse werkwijze. Wat is GPU computing en hoe wordt dat verschil verkregen? Met name de schermafdrukken met de instellingen voor het kiezen van de rekenmotor zijn interessant.

Five steps to GPU power in Vegas Pro


Waarom Furryball?

GPU renderer

Raytraced Physically-based full global illumination ray tracing

What is GPU rendering?

GPU rendering means, that there is used the power of GPU (Graphics card) for rendering, instead of CPU (processor).

Conventional CPUs (Intel, AMD) didn't make such dramatic progress, like GPUs in last 10 years. What is GPU rendering - NVIDIA article

De beste grafische kaarten voor gebruik met Furryball

Systeem vereisten

CUDA vs. OpenCL

OpenCL vs. CUDA: Which has better application support? GPGPU/GPU Acceleration real world face off

AMD vs. Nvidia:

Should I go with OpenCL or CUDA? Which will perform best with my applications?

GPU Rendering vs. CPU Rendering

A method to compare render times with empirical benchmarks

Posted on October 2, 2014 by Joe Pizzini


A growing number of customers are using GPUs for big data analytics to make better, real-time business decisions. This page highlights customer use cases and their techniques in big data analytics, such as machine learning, search, and sorting.

What is the future of 3D rendering?

Rendering is changing:

the traditional CPU render model is being challenged, and Mike Griggs suggests that it's time to re-examine your render solutions.

What is GPU rendering

Voorzien van de bekende Mythbusters commercial die door NVIDIA is betaald.

This is the real power of GPU. Never possible on CPU in realtime...

PhysX FleX is a particle based simulation technique for real-time visual effects.

Introducing Indigo's pure GPU rendering

One of the worst-kept secrets in the Indigo community is the forthcoming pure GPU rendering support, and today we'd like to show you some exciting first results!

Free Deep Learning course from NVIDIA

Deep learning is a rapidly growing segment of artificial intelligence. It is increasingly used to deliver near-human level accuracy for image classification, voice recognition, natural language processing, sentiment analysis, recommendation engines, and more. Applications areas include facial recognition, scene detection, advanced medical and pharmaceutical research, and autonomous, self-driving vehicles.

Boxx zijn visie op GPU rendering versus CPU rendering.

CentiLeo renders video shots for big scenes on a laptop


SketchUp Pro scene klaar voor renderen

Dezelfde SketchUp Pro scene in Octane Render (een allereeste versie in de begintijd van Octane, hieronder).

Value-based GPU build for Octane Render

February 9, 2015/in Uncategorized /by @tg

It's easy to go to Your favorite shop (or e-shop) & buy the most expensive product willing that it will perform the best. If You done so, well.., I have bad news for You. Reality is somehow a bit more complicated & pricing is usually based on different performance metrics.

Rob Bredow on GPU-accelerated rendering

Jan 7, 2011 00:00 am Comments

Sony Pictures Imageworks' Rob Bredow explores the pros and cons of GPU-accelerated rendering in real-world production pipelines.

Videos tagged with gpu render

V-Ray 2.0 instelling vor CPU - OpenCL of CUDA met 3ds Max interface.

Should one invest in CPU or in GPU for CGI rendering?

Jae Alexis Lee, This is what I actually went to school for!

It depends entirely on your workflow.

A trip through the Graphics Pipeline 2011: Index

July 9, 2011


About the D3D / OpenGL graphics pipelines as actually implemented by GPU's. A lot of this is well known among graphics programmers, and there's tons of papers on various bits and pieces of it, but one bit I've been annoyed with is that while there’s both broad overviews and very detailed information on individual components, there’s not much in between, and what little there is is mostly out of date.
* Part 1: Introduction; the Software stack.
* Part 2: GPU memory architecture and the Command Processor.
* Part 3: 3D pipeline overview, vertex processing.
* Part 4: Texture samplers.
* Part 5: Primitive Assembly, Clip / Cull, Projection, and Viewport transform.
* Part 6: (Triangle) rasterization and setup.
* Part 7: Z / Stencil processing, 3 different ways.
* Part 8: Pixel processing fork phase
* Part 9: Pixel processing join phase
* Part 10: Geometry Shaders.
* Part 11: Stream-Out.
* Part 12: Tessellation.
* Part 13: Compute Shaders.

Low-level Thinking in High-Level Shading languages

Emil Persson

Game developers conference 2013

Bunkspeed interface

Onderaan de pagina een groot aantal links over de werking van AMD, NIVIDIA en Intel GPU's.

GPU Rendering vs. CPU Rendering, Empirical approach for comparing render times

Preview Book

GPU PRO 3: Advanced Rendering Techniques

febr. 12, 2012

How-To Geek

How Do The CPU and GPU Interact to Render Computer Graphics?

OpenGL rendering Pipeline

Computer Graphics World

Boxx antwoord op GPU rendering


The Most Advanced Workstation For GPU Rendering

Tot 5 grafische kaarten bij elkaar met de multi-GPU 3DBoxx 8950. !

Opposite Renderer

GPU Photon Mapping

Master's thesis project at NTNU.

Progressive Photon Mapping with GPU. An introduction to ray-tracing, photon mapping, participating media, as well as details on implementation and results are available in the full thesis.

Download this thesis (PDF)

GPU benchmark

2015 GPU rankings

The CPU is faster and more powerful then GPU?

12 pages PDF

PDF download

The Art of Rendering (updated)

By Mike Seymour

April 10, 2012

Renderers are easy to write in the abstract, as perhaps a university project, but to work in production environments is extremely difficult. Arnold, by Solid Angle, is some 200,000 lines of highly optimized C++ code, and it is considered a very direct implementation without a lot of hacks or tricks. Production requirements in terms of rendertime and scene complexity are staggering. And the problem is not just contained to final render time, as Arnold founder Marcos Fajardo pointed out at Siggraph 2010

Final render CPU time might cost $ 0.10 per hour, but artist time is closer to $ 40 an hour, so interactivity is also vital.

GPU vs CPU rendering Talk by Luxology


Thursday, August 26th, 2010 | Posted by Leonard Teo

Het betreft een van de eerste versies van Octane. Bekijk de website voor de nieuwste uitvoering.

Octane Render GPU-based rendering

Monday, May 17th, 2010 | Posted by Leonard Teo

GPU box & Octane

16 GPU's (8 x GeForce GTX 690 + CPUBOx = 1:01
4 GPU's (2 x GeForce GTX 690) + GPUBox = 4:07
4 GPU's (2 x GeForce GTX 690) - native = 4:05

Intel i7 3,820 GHz.

Published on Aug 12, 2014

The video compares native rendering in Octane on 4 GPUs (2 x GeForce GTX 690) to rendering on 4 and 16 GPUs virtualized with GPUBox Artist. Using Octane Render within the GPUBox infrastructure means nearly native performance and great scalability. You can find out more in the post on our blog:

CUDA Based render farms

Renegatt Software has released GPUBox Artist, a new tool for setting up and administering CUDA-based GPU render farms. The software, which is still officially in beta, works with both Blender and Octane Render.

(Full disclosure: it came out earlier this month, but we haven’t seen much coverage, and it's an interesting tool.)

A simple GPU virtualization solution

GPUBox Artist ‚Äúseparates application and operating systems from the [physical] GPUs‚, enabling a renderer to access GPUs in servers or other artists machines across a network.

To access a set number of GPUs, the user enters a simple terminal command ($ gpubox add n to access n GPUs, or $ gpubox drop all to stop using them again), or uses a web-based client.

CUDA renderers on the user’s machine can then access GPUs on the network as if they were installed locally. Several users can share the same GPU; and users can access GPUs from virtual machines.

Simple as that. The software also provides administrators with basic security and network monitoring options.


Visualization and HPC Solutions

Long coffee breaks are a thing of the past using GPU-Xpander! Part I

You've just completed weeks of conceptualizing an image or product design, building a 3D model with hundreds of thousands of polygons (or more), you hit the render button in your CPU-bound rendering application, and.......wait......FOREVER!!

A Cubix customer recently sent me some interesting tests results he got from doing a comparison of software, CPU-based vs hardware, GPU-based rendering. He used the same 3D model consisting of hundreds of thousands of polygons. He compared various resolutions ranging from 1280 x 720 to 1920 x 1080, and 1500 to 5000 iterations.

Using just his dual Xeon 3.33Ghz (12 cores) workstation and no GPUs, he got a 1280 x 720 @ 1500 iteration render in just over 3.5 hours, and a 1920 x 1280 @ 5000 iterations render in 36 hours.

Using 5 NVIDIA Fermi-based graphics cards, four of them hosted inside a Cubix GPU-Xpander Desktop 4 system, 1 in the dual Xeon workstation, he did the 1280 x 720 @1500 iteration render in just 25 minutes! The 1920 x 1080 @ 5000 iteration render took only 4.5 hours!


Multiple GPU-Xpanders on one PC?

Yes, of course you can run multiple GPU-Xpanders from the same motherboard! Marco Podrini from Studio Podrini in Italy just sent me a picture from his office showing his PC workstation attached to dual GPU-Xpanders. He uses his GPU-Xpanders for GPU-accelerated rendering, among other applications.

Why did he do this? He wanted the same things most other Cubix GPU-Xpander customers want: PC to run cooler, separate the GPUs from system, more slots & watts, flexible I/O configurations.

Workstation specialst GPU rendering

"The Worlds Fastest, Heterogeneous Rendering System"

Renderstream van TYAN

8 CUDA cards zonder de noodzaak voor een speciale bios software?


Server barrebones

TYAN, an industry-leading server platform design manufacturer and subsidiary of MITAC Computing Technology Corporation, presents its cutting-edge platforms validated with NVIDIA Tesla K80 dual-GPU accelerators.

4U Rackmount, FT72-B7015_V2.1 is uitgerust met:

Voeding van 2 x 1200 Watt en moederbord met 8 stuks PCI E gen. 2 x 16 sloten voor grafische kaarten. En twee Intel Xeon Processors uit de 5500 series.

Met de keuze van de NVIDIA kaart dient u met generatie 2 dus rekening te houden in uw keuze van de juiste kaart. Standaard zijn de moderne kaarten allemaal gen. 3.0.

Of informeer bij de fabrikant van deze rackmount of dat inmiddels aangepast is.