Slashdot Mirror
Audio Processing on Your Graphics Card?
edsarkiss writes "BionicFX has announced Audio Video EXchange (AVEX), a technology that transforms real-time audio into video and performs audio effect processing on the GPU of your NVIDIA 3D video card, the latest of which are apparently capable of more than 40 gigaflops of processing power compared to less than 6 gigaflops on Intel and AMD CPUs." Another reader points out a story on Tom's Hardware.
This kind of stuff has been talked about and done in the research community for quite some time now. See http://www.gpgpu.org. While audio is an interesting idea, FFT's and Genomics are already running on GPUs Yes, GPU's can be fast, but they can also be a pain to program. Take a look at the Stanford Brook for GPU's project for a nice elegant way to program for GPUs. http://brook.sourceforce.net
Emmm, what about this, for example?
I'm just surprised that nobody thought of doing it sooner.
It's easy to be surprised when you're wrong: BrookGPU: General Purpose Programming on GPUs December 2003.
Anytime there is an article talking about the power of your graphics card's GPU or the phenomenal processing power of DSPs, the discussion is always inundated with people asking "Hey why aren't we using these instead of our regular slow processors!", thinking they've come up with some sort of brilliant idea. For the thousandth time, people, things just don't work that way. DSPs achieve their high processing speeds by being very good at a few select things, but not really being general purpose devices. If you want to know more of the specific details, do a google search, there's a ton of information about DSPs on the web and I'm sure there are plenty of pages that explicitly address the difference between CPUS, GPUs and DSPs.
I believe slashdot has already covered stories about using the programmable pipeline on mondern gpus for non-graphics functions. They're built to crunch vectors/math. Why not?
Find Nearby Indie Events
People are doing extremely interesting things with modern graphics hardware, including fluid dynamics simulation, cloud simulation and multiplication of large matrices.
A good site for information on it is www.gpgpu.org, where there are perhaps 200 different projects related to general purpose GFX card use.
As the capabilities of graphics cards expand and become more esoteric, perhaps game developers will begin to eschew the use of certain graphics featuers in favor of using those parts of the pipeline to perform generic calculations, such as physics.
Perhaps there are also ways of performing such calculations and using the results as decorative graphics, ie when we're showing decorative ripples on water, perhaps those ripples are artifacts of some calculation that is being used elsewhere in the game.
Intolerance for ambiguity is the mark of the authoritarian personality.
This phenomenon is commonly known as the "Wheel of Reincarnation". Diverting functionality to specialized components, and then folding it back onto the CPU has been going on since the 60s.
A more detailed description of the WoR is available here.
Tom Rokicki computed the Game of Life using Amiga's Blitter. March 17, 1987 UseNet post.
And another post:
How can the price range be so slow when the processing power is claimed to be so many times faster than Intel chips?
First, silicon area doesn't necessarily mean performance. The whole reason that IBM, AMD and Intel are building multi-core chips is that so much of the area in a moden microprocessor is spent in workarounds for different structural hazards rather than in real work. The GPUs are huge because they are parallel mathematical computation engines. On a FLOP per sq. mm basis, they are a LOT more efficient than a single core CPU could hope to be.
As WIAKywbfatw points out, GPUs became more powerful than CPUs (on a FLOP basis) a decade or more ago. This was the whole reason Intel created the AGP port - to prevent the GPU from becoming the center of the the computer (it was a huge threat to their business).
Today, silicon is more and more about customization... on a FLOP basis, the chips in HD digital TVs have nearly the performance of the latest P4 - but at MUCH less cost... because they are less flexible (a LOT less flexible). Their design is to optimize single precision floating point performance... You can't use that CPU power for a long-running simulation ("scientific computing") - only for graphics; where single precision is still orders of magnitude more precise than the monitor can display.
I work in the video games industry. Using graphics processor for audio is not new. The Nintendo 64 had a "Reality-Engine" graphics coprocessor that also processed sound by uploading new microcode.
If you think about it, things like bilinear/trilinear filtering are perfect for resampling, graphic blendops like add/subtract/modulate are great for audio mixing and can be done with even older fixed function hardware and bit of programming effort. The programmability of new hardware with pixel and vertex shaders improves the generic applications of the GPU by orders of magnitude and allows significantly more non-graphic algorithms to be implemented.
As mentioned, the entire General Purpose GPU (GPGPU or GP2U) community has been on this a while. They recently put on a SIGGRAPH 2004 Course and there is the GPU Gems book which as some GPGPU in it.
The power of GPUs comes from their specialization at performing certain computational tasks, tasks such as manipulation of large matrices and bitwise operations on huge datasets. These tasks are not limited to pixelpushing, but many kinds of task (such as audio processing) can be visualized as operations on graphical bitmaps. For instance, a reverb filter is very similar to a directional blur in photoshop.
There are also many optimizing tricks involving GPUs that may lend themselves to certain tasks more than others. For instance, feeding data to the card is expensive, so you need to do it in chunks of a certain size. Read the gpgpu.org site and you'll see what I mean-- there are many explanations on there of why GPUs are good at some things and not so good at others.
Intolerance for ambiguity is the mark of the authoritarian personality.
The UAD-1 card is actually a video accelerator card. The VST plugins written for it are just wrappers for specific code performed on the card.
It's not possible with AGP, but have no fear. PCI Express to the rescue!
Its been thought of before, however the performance of a GPU compared to a CPU purely by clock speed or mips is sort of comparing apples to oranges...
GPUs are special purpose.. CPU's are not..
---- Booth was a patriot ----
You forgot a qualifier- massively iterative and embarrassingly parallel operations. The current generation of graphics chips are pretty much 2 or 3 previous generation designs shoved into the same core; the Geforce 6800 and ATI X800 both have 6 vertex pipelines and 16 pixel pipelines.
Given that OpenAL is backed by sound card manufacturers, I wonder if they would ever concede to using GPUs to accelerate 3-D sound. I hope that the apparent conflict of interest doesn't hinder progress, if GPUs can really make a difference.
OpenAL is the one cross-platform audio API I've tried that actually _works_, while the other cross-platform options seem to either be stagnant, incomplete, just plain garbage, or so lacking in documentation that no mere mortal could figure them out. Here's to hoping that OpenAL and cross-platform audio on UNIX keeps getting better and better, because we really do need it.
-- "Makes Little Debbie look like a pile of puke!" - Moe Szyslak
As WIAKywbfatw points out, GPUs became more powerful than CPUs (on a FLOP basis) a decade or more ago. This was the whole reason Intel created the AGP port - to prevent the GPU from becoming the center of the the computer (it was a huge threat to their business).
nVidea used the term GPU to refer to its fixed function T&L capable NV10 chip, which was released on August 31, 1999 as the GeForce 256.
The AGP 1.0 standard dates to 1996, and it was intended to provide fast bandwidth for textures and video. The video cards it fed were generally incapable of running shaders, geometry transformations, or lighting calculations.
Programable floating point units are a relatively recent development-- the functionality was introduced to the mass market when the Radeon 9700 and GeForceFX 5700 were released.
I think that's why he sent the jobs to the printer to be processed. He wrote the "program" in postscript, and the printer, with its specialized processor did all the work of generating the graphics and calculating the force and motion vectors. Postscript is great for doing math on vectors and it can output vectors to be drawn, or just output their parameters as text values. He could get the printer to process a simulation in 30 minutes that would take a day running on his local box.
The printers had the best processors available for his work. Of course, he was the first of our profs to have a linux box on his desktop, and the first to do parallel processing on the several sun sparc 2 workstations we had.
He was always pretty clever at using the computing resources at hand!
PDF is rather simpler than Postscript. You can program in Postscript-- whereas you can merely draw with PDF.
I really can't understand this hostility towards PDF. Some renderers may be quite slow, but Apple has shown that it's perfectly possible to write a efficient Preview application. Perhaps someday Adobe will step up to the challenge.