The most recent Catalyst 8.9 Linux driver enables use of multiple FireGL and FirePro cards to act as a single X server that spans multiple displays, each using OpenGL acceleration. Our Siggraph 2008 gallery video showed an example of this with the open source FlightGear running across 4 displays (each at 1920 X 1200 pixels) driven by 2 FireGL v7600 accelerators under Linux.
I received an email in response to our gallery posting - upping the ante! The 10 minute video shows four FireGL v5600 cars driving eight OpenGL-accelerated monitors running the open-source FlightGear software. Impressive!
Tech Specs
- AMD Phenom 9550
- ASuS M3A32-MPV Deluxe (AMD 790FX)
- 2 GB RAM
- 4 x FireGL 5600
- 8 Monitors
- Ubuntu 8.04
- ATI Catalyst Linux Driver 8.9
- Flightgear CVS (20080720)
In January 2008, it was discovered that licenses for the SGI OpenGL Sample Implementation and the GLX API / extensions that are at the heart of GNU/Linux 3D applications (GLX is the glue connecting OpenGL and the X Window System), were not actually free. The original SGI Free License B and the GLX Public License were permissive, but contained terms which which were a significant problem for the free software community. But SGI has come through and released a new version of the SGI Free Software License B, which now mirrors the free X11 license used by X.Org. Both Mesa and the X.org Project can continue to use this code in free software distributions of GNU/Linux. This is great news for the open source 3D scientific, CAD and design communities.
In Part 1 - Cost, I wrote about the inevitably of DisplayPort becoming the dominant display interface standard for PCs and handheld devices, if for no other reasons than cost:
DisplayPort avoids the $10k/year license fee of HDMI
DisplayPort direct-drive technology eliminates the cost for additional circuitry in computer displays
In Part 2 - Performance and Design, I reviewed some of the performance advantages of DisplayPort over DVI and to some degree, HDMI:
DisplayPort uses direct-drive technology, enabling ultra-thin displays and a setting a common standard for laptop and stand-alone displays
DisplayPort can drive 30-bit (billions) of colors at high-resolutions and high refresh rates
DisplayPort uses smaller, latching connectors and can handle longer cables
DisplayPort offers great flexibility in handling, video, audio and data
In this third installment I want to look at DisplayPort in terms of its potential for future digital devices relative to DVI and HDMI.
Futures
Unlike DVI or HDMI, both of which include legacy technology for CRTs, DisplayPort was designed specifically to handle today’s digital displays and to be able to adapt to new features in the displays and devices of the future.
The 45-nm Holy Grail of Chip Size
Chip makers from Intel to AMD to Nvidia are striving to reduce die size to 45-nm and smaller in order to reduce power consumption and increase transistor count. However, a 45-nm process imposes a technical limitation of 2.5V maximum for I/O transistors. HDMI and DVI both use TMDS, which requires 3.6V when running high-speed signals (up to 5.25V for low-speed sideband signals).
The only way around this limitation for HDMI and DVI is to add proprietary, custom circuitry. This translates into increased complexity and increased cost.
DisplayPort in contrast, requires less than 2V for high speed signals, so it can be integrated with a standard 45-nm process - no custom circuitry, no added size, and no additional costs.
Picture-in-Picture and Daisy Chained Displays
As already discussed in Part 2 - Performance and Design, DisplayPort was designed as all digital and uses micro packets to bundle audio, video, and data information. This micro-packet protocol is designed to support more than one audio or video stream, as well as data - all over a single cable. The current DP specification allows for up to six 1080i streams or three 1080p streams. So in the future you should expect to see things like Picture-in-Picture or daisy chained-monitors without additional cables or circuitry.
USB, Webcams, and Touch-Sensitive Displays
DisplayPort not only offers a scalable data channel, it offer a bi-directional scalable data channel. This means that future implementations will be able to support microphones, webcams, USB hubs, or touch-sensitivity built right into the display - without additional cabling and circuitry (this would be especially valuable on laptops where physical space is at a premium).
Dongles and Backwards Compatibility
DisplayPort may be the future, but since a lot of us are dealing with existing displays, laptops and consumer-electronics devices, there will be a need for adaptor cables or dongles to bridge between HDMI/ DVI and DisplayPort.
DisplayPort already offer pass-through support for HDMI signals. So using the appropriate dongle you can hook up your DVD player or cable box to a DisplayPort monitor. Coming later this year, expect to see adaptors that will allow you to connect your DisplayPort video card, to an HDMI or DVI device.
Final Score Card
So let me update my score card:
Feature
DVI
HDMI
DisplayPort
Price
+1
0
+2
Sleeker Displays
0
0
+2
Smaller Connectors
0
+2
+2
Resolution & pixel depth
+1
+2
+2
Long cables
0
+2
+2
Auto-tuning
0
?
+2
Audio
0
+2
+1
Smaller Chips
0
0
+2
Picture-in-Picture
0
+1
+2
Webcams & Touch Displays
0
0
+2
Compatibility
+1
+1
+1
While there is not a clear winner between HDMI and DisplayPort for today's market, in the near future the demands for the, cost benefits, performance, and scalable spec of DisplayPort will become more commanding. The FirePro line from AMD, as well as 30-bit color monitors from Dell and HP, are strongly embracing DisplayPort (but still hedging with at least one legacy DVI port) and I expect to see more card and display vendors move in this same direction next year.
Addendum: For the complete 3-part series on DisplayPort see:
For a while, it was feeling like the Radeon HD 4870 had taken the lead in advanced hardware over the FireGL line (CAD-optimized drivers and RAM aside). But today AMD updated its flagship FirePro workstation accelerator to offer the same technology that has made news in the Radeon HD 4800 series: 800 unified shader processors and GDDR5 memory. But the FirePro v8700 also brings 1GB video memory and optimized drivers for CAD, digital video editing, and 3D visualization. It also adds twin Display Port connectors for the latest generation of hi-rez, hi-refresh, 30-bit color rendering displays as well as a dual-link DVI connector for older monitors. AMD’s press release claims a 40% performance gain over previous FireGLs.
The SIGGRAPH 2008 video gallery is our way to bring a bit of the AMD booth to you. Mach Studio was on display, showing real-time CG (reducing production cycles that could take days or weeks down to hours, minutes or even seconds) using the FirePro GPU. But we live camcorder demo from the show did not come out clear so Studio GPU created a video for FireUser.com. We have added this video demo to the gallery.
As reported in the Siggraph 2008 blog by Pat Howk, Autodesk is making a big push to take a market lead in stereoscopic 3D for media & enertainment.
Their latest announcements at IBC further support this. Lustre 2009 is their digital color grading system (color grading is a process for altering or enhancing the color of movie or TV images). Lustre’s two big updates:
The entire array of grading, previewing and rendering tools can be used for Stereoscopic 3D material
Real-time playback of primary and secondary color grading capabilities for multiple secondary layers are GPU-accelerated. FX plug-ins (blur, noise, etc) are also GPU-accelerated
The Autodesk support for stereo 3D in their film and game development software moves them to the forefront of the market. Stereo 3D as part of the mainstream for CAD and visualization will most likely be next as display vendors begin to release their auto-stereoscopic displays. I am also guessing that there will be a lot more work on 3D stereo drivers from AMD and Nvidia as this market moves from infancy to adolescence.
While this may be old news for the digital content creators, I see a lot of industrial designers and engineers neglect ambient occlusion in their renderings. This is unfortunate, because studies have shown that this property of light allows us to perceive geometry better than direct lighting (Depth discrimination from shading under diffuse lighting, by M.S. Langer and H. H. Buelthoff). Communicating geometry is the intent of rendering for design ultimately.
Ambient occlusion is a decrease in light due to occlusion. It is the shading if the object was lit evenly from every direction, similar to an overcast day. Soft shadows form in cracks, crevices and corners because the surrounding geometries don’t let as much light in at those points. If you look at the room you are in right now, you can notice that area of the wall closer to the corner of the ceiling is slightly darker than the rest of the wall with the same light. This is because the wall blocks, or occludes, light from more and more directions as it moves closer to the corner.
While this is a subtle effect, it is one of the primary visual cues our minds look for to discern if something is real. It also shows off geometry very well. Thinking about it in your sketches or hand rendering can make your art much more realistic. As for 3D, your rendering program probably can do it, but it may require turning on Indirect Illumination, Global Illumination, or Ray Tracing. (For example, see Ambient Occlusion In Any Scene for how to enable the effect in SolidWorks). It will increase your render times, but you’ve got the horsepower so might as well use it!
SIGGRAPH 2008 was an amazing show for anyone interested in 3D and visualization. This video gallery is our way to bring a bit of the AMD booth to you.
Video demos we have include:
- Real-time visualization of 350 million polygon CAD model
- Linux flight simulation across four 1920 X 1200 displays
- Complex fluids simulation using the GPU
- Real-time photorealistic visualization for marketing
- Dynamically-generated level-of-detail and collision detection
- Real-time medical CT scan volume reconstruction.
No CAD or DCC professional likes waiting for their system to update the screen. So a principal goal for any modern 3D rendering application is real-time rendering and interactive editing. One approach to achieve this level of performance, is to use sorting, culling or LOD to limit the number of polygons that need to be rendered. Another approach covered in this article and video, is to move raw geometry handling from the CPU, to the the GPU using OpenGL’s Vertex Buffer Objects (VBO).
VBO is a way for the workstation graphics card to store the geometry in the card’s framebuffer and even update it without having to push the geometry back down the bus to the CPU to calculate changes. This more efficient use of framebuffer memory reduces the instances the GPU is waiting for the CPU and and significantly accelerates the performance of things like large model rotations. Read article →
There has been much controversy over the direction the Khronos Group/OpenGL ARB has chosen for the next major version of OpenGL. After testing an approach that would have a drastic effect on the API, requiring complete OpenGL application rewrites and not introducing any of the long awaited features modern GPUs are capable of, the choice was made to give programmers what they are really waiting for. And that’s new features now. GL 3.0 takes two important steps to moving open standard graphics forward in a major way. The first is to provide core and ARB extension access to the new and exciting capabilities of hardware. The second is to create a roadmap that allows developers to see what parts of core specifications will be going away in the future, also providing the OpenGL ARB with a way to introduce new features faster.
Over the last few years graphics hardware has made great strides forward. Different vendors have exposed home-grown extensions to give users access to hardware, but vendor extensions vary between vendors and are not a stable approach to supporting new hardware. GL3.0 brings these new features under one roof, defining one common and accepted way that all vendors will implement. Now all GL application programmers can get access to things like float color/texture/depth buffers, integer formats, conditional rendering, framebuffer objects, transform feedback, vertex array objects, half-float data types (vertex/pixel), and so much more. With these new features all developers have the tools to add new eye candy and much better optimize render algorithms and performance.
Many of the new features of OpenGL3.0 provide mechanisms to increase the efficiency and speed of today’s complex scenes. Conditional rendering allows an app to discard geometry that would be occluded during normal rendering. Half float formats can help drastically compress vertex data sets. Vertex array objects make setting up rendering much easier and less error prone. Map buffer range support allows a small portion of a buffer to be mapped, even while it is rendered from, no more GPU stalls required. There are also quite few additions to enhance rendering flexibility. Framebuffer objects provide a fast and simple way to accomplish off-screen rendering. Transform feedback opens the door to a whole new set of complex multi-pass rendering and geometry generation algorithms previously impossible. Integer-in-shader support allows for much more flexible and natural shader code. Several other important features such as instanced rendering and geometry shaders have now been given ARB extension status as well.
By introducing the new deprecation model, the ARB has created a way to signal what will be removed in future revisions. This provides enough time for all developers to move code-bases to newer and better methods. Future versions of GL will remove fixed-function rendering, color index mode, immediate mode, client vertex arrays, and other seldom used portions of older specs. This helps to keep the spec lean and mean, also allowing hardware vendors to better optimize performance and maintain quality. A way for OpenGL to gracefully move forward has been long missing. With the most recent changes, OpenGL now has the tools to keep open standard graphics current and useful for many different flavors of 3D applications.
FireUser.com is a community resource for CAD, visualization, 3D, video and engineering professionals to learn about the latest acceleration and display technologies and news with a focus on the AMD FirePro workstation graphics line.
