Symphony + Nvidia FX 5800 + Logic + Snow Leopard?

[nativeaudio]

Apple's next OS, Snow Leopard (OSX 10.6) will support GPGPU, and OpenCL and Grand Central.

Snow Leopard further extends support for modern hardware with Open Computing Language (OpenCL), which lets any application tap into the vast gigaflops of GPU computing power previously available only to graphics applications. OpenCL is based on the C programming language and has been proposed as an open standard.

GPGPU (General Purpose Graphics Processing Unit computing) has been used for audio signal processing.

A few days ago, NVIDIA announced a new graphics board (two PCI slots wide), costing $3,500, with 4gbytes RAM and a throughput of 102 GB per second (!).

Who needs a $3,500 graphics card? This is where it gets interesting, because this 240-core card is designed for professional users that want to "tap the parallel processing power of graphics chips to crunch massive amounts of data without overloading the microprocessor inside a computer".

Sounds brilliant, especially for people who work with Final Cut Pro perhaps, but if this card can do audio processing as well, or at least extremely fast pre-rendering (freeze-on-the-fly" - it apparently allows "extreme floating point calculations") couldn't this card possibly turn everything upside down for the non-native audio world?

Note that these ideas aren't really new. Here's a thread called Nvidia cards will process audio soon from 2004.

[SoundEng1]

WOW!

Snow Leopard with support for GPGPU, OpenCL,& GC, is going to be a Audio Beast!

[nativeaudio]

Originally Posted by SoundEng1

WOW!

Snow Leopard with support for GPGPU, OpenCL,& GC, is going to be a Audio Beast!

Possibly, but can anyone with more knowledge about these technologies chime in with some info about whether what I wrote in my initial post The assumption that this could have a big impact on how people work with audio) is totally nonsense, or it it actually makes at least a little sense?

I would love to hear comments from someone with more knowledge then myself about the technical aspect of using graphics boards for real time audio processing. All I know is that 'it has been done'.

[hmurchison]

I'm certainly not an expert but I'm hoping that OpenCL will be leveraged for some fantastic audio applications. Even if the latency precludes using OpenCL for realtime applications I could see it being used to do non-realtime processing.

More processing power leveraged by creative ingenuity will yield some amazing results.

I see a few big drivers in computer technology that will affect us all.

1. The rise of programmable GPU-

We'll laugh out loud one day when we remember how GPU were "just" used for graphics display.

2. Many-core CPU and better threading tools-

Soon we'll have 8- cores on a chip with multithreading so that they can appear as 16 processors to your OS. Developers should be somewhat insulated from the vagaries of programming these threads with tools like Grand Central.

3. Chip fusion-

We're going to have CPU and GPU on the same die soon and Intel's working on a new SIMD engine named AVX for their 32nm Sandy Bridge chip. 256-bit performance will double today's processing power.

4. SSD creating a storage revolution-

Luckily we audio guys don't have the storage requirements that the video guys do. This means SSD drives will be viable for us long before they become affordable for the needs of video. In a few years we'll see the size of today's drives quadruple and the speed go up.


It's easier to see the computer of 2014 now.

8 5Ghz cores with multithreading
1 Terabyte of SSD storage
4GB GPU
Blu-ray with 100GB optical disc support
USB 3.0 and Firewire 3200

And these computers will use less power than today's computers.

[nativeaudio]

I Googled, and found some links that seem to confirm what I hoped to see, namely that audio was one of the areas that would be particularly interesting in terms of what OpenCL and GPGPU can deliver.

http://www.exampler.com/blog/2008/06...-contingency/:

Actually, we’re not as much “taking it back” as we are finding that the massive-parallel architecture is well suited for tasks other than graphics, such as physics or audio; with GPGPU, we can do things that are basically impossible on the CPU, no matter if that’s a 10yr old CPU or tomorrow’s monster.

Windows 7 to feature GPU acceleration like Apple's Snow Leopard? - Engadget

Both nVidia and AMD/ATI have been developing and marketing SDK solutions that allow programmers to take advantage of the massive computational capabilities of GPUs for data-parallel processing (think audio/video encoding/decoding/processing, graphics manipulation, molecular modeling, scientific computing and simulations, Oil/Gas geo-engineering, visualization, etc.).

OpenCL: What you need to know | Macworld

Any app could in theory make use of the technology. As another poster commented, it's a bit like the MP macs back in the Power Mac 9500 / 9600 days, except back then apps needed to be specifically coded / re-written / re-designed to be multi-processor aware.

The same is true today on Mac OS X (apps need to be split up to divvy out the work across processors), except these days, Mac OS X is able to load balance threads / processes for you (whereas in OS 8 / 9 it was down to you).

Consequently even apps that aren't split up / can't be re-designed will still benefit from Open CL, as the kernel will be sending processes that can make use of the GPUs onto them. That in turn will free up the CPU for things that remain CPU-based. Initially that will be "most apps" of course, but over time (and depending how easy it will be for developers who aren't used to multi-threading / multi-processing to learn to use Open CL), apps will potentially be evenly distributed over the available technology in the machine.

In terms of apps that can make direct use of Open CL, anything that processes large swathes of data could be a candidate. It doesn't have to be graphical apps (the fact it's a graphics card doing the work is irrelevant for Open CL apps)... any audio app could in theory do the work, as could databases etc.

OpenCL - Wikipedia, the free encyclopedia

OpenCL (Open Computing Language) is a language for programming heterogeneous data and task parallel computing across GPUs and CPUs. It was created by Apple in cooperation with others, and is based on C99.
The purpose is to recall OpenGL and OpenAL, which are open industry standards for 3D graphics and computer audio respectively, to extend the power of the GPU beyond graphics (GPGPU).

[nativeaudio]

Some more links:

Rumor: Nvidia designing GPGPUs for Apple computers

A GPGPU, according to the website’s definition, is a unit designed for “general purpose computing on graphical units.” The new unique hybrid unit is based on Nvidia’s architecture called CUDA, for compute unified device architecture.

What the CUDA design allows for is highly intensive number crunching, scientific equation handling, physics rendering, and standard video and audio processing. Essentially, it will allow your already powerful computer to be even more powerful.

That means you probably won’t be seeing the new units built into your Macbook Air anytime soon; if the rumor is true, the units will be for professional machines, mainly the Mac Pro. Just like having an 8-core processor, you can bet your britches you’ll be spending more than pennies on the new unit; Nvidia has a GPGPU developed right now called the Tesla that runs for $1500.

Though no exact numbers were given as to what kind of audio and graphical improvements could be expected, in regards to scientific calculations, the Tesla unit has performed from 45x to 415x faster than other processors while processing hospital MRIs. Essentially, the GPGPU can reduce the work of a computer cluster to a single machine; talk about an increase in productivity.

Efficient 3D Audio Processing on the GPU

BionicFX NVIDIA GPU Audio Effect Processor

BionicFX Harnesses Power of Graphics Cards for Audio Processing
Programming uses GPU as Audio Effect Processor

BionicFX announced a technology for music production that turns NVIDIA video cards into audio effects processors. Audio Video Exchange (AVEX) converts digital audio into graphics data, and then performs effect calculations using the 3D architecture of the GPU. The latest video cards from NVIDIA are capable of more than 40 gigaflops of processing power compared to less than 6 gigaflops on Intel and AMD CPUs.

GPU audio processing: Two VST plugins, chorus and delay


GPGPU

Graphics-based Acoustic Simulations

Physically correct acoustic simulations for complex and dynamic environments remain a difficult and computationally extensive task. Graphics hardware is here used for the simulation of sound wave propagation. Two different methods have been implemented, of which one uses ray tracing techniques, while the other is based on difference equations and waveguide meshes. Both techniques can efficiently be implemented within a real-time environment by concentrating on the similarities for sound and light wave propagation, and by exploiting the possibilities of using graphics hardware for non-graphics computations. Applications are discussed for real-time room acoustics, virtual reality as well as for virtual HRIR measurements based on polygonal meshes.

(Ray Acoustics using Computer Graphics Technology. Niklas Röber, Ulrich Kaminski, and Maic Masuch. Proceedings of DAFx 2007.)
(Waveguide-based Room Acoustics through Graphics Hardware. Niklas Röber, Martin Spindler, and Maic Masuch. Proceedings of ICMC 2006.)
Posted: 05 Nov 2007 [GPGPU /Audio and Signal Processing]

Audio and the Graphics Processing Unit

From the abstract: In recent years, the development of programmable graphics pipelines has placed the power of parallel computation in the hands of consumers. Systems developers are now paying attention to the general purpose computational ability of these graphics processor units, or GPUs, and are using them in novel ways. This paper examines using pixel shaders for executing audio algorithms. We compare GPU performance to CPU performance, discuss problems encountered, and suggest new directions for supporting the needs of the audio community. Source code is also available. (Audio and the Graphics Processing Unit", by Sean Whalen)
Posted: 16 May 2005 [GPGPU /Audio and Signal Processing]

Room Acoustics Computation on Graphics Hardware

In this masters thesis by Marcin Jdrzejewski, ray tracing is implemented on the GPU to accelerate computation of sound paths between sound sources and receivers. Each ray averages 16-20 wall reflections, and those rays that intersect a sphere approximating the receiver are included in an echogram that is used in the auralization process. Typically 4096 rays are used, but the application can run in real time with up to 64K rays. A demo application, article and some movies can be downloaded from the following link. (Computation of room acoustics on a GPU.)
Posted: 06 Oct 2004 [GPGPU /Audio and Signal Processing]

BionicFX uses GPU as Powerful Audio Effect Processor


"BionicFX announces a revolutionary technology for music production that turns NVIDIA video cards into audio effects processors. Audio Video Exchange (AVEX) converts digital audio into graphics data, and then performs effect calculations using the 3D architecture of the GPU. The latest video cards from NVIDIA are capable of more than 40 gigaflops of processing power compared to less than 6 gigaflops on Intel and AMD CPUs. AVEX represents a major technological achievement that allows music hobbyists and professional artists to run studio quality audio effects at high sample rates on their desktop computer. (Press Release: "Revolutionary Programming and Innovation uses GPU as Powerful Audio Effect Processor")
Posted: 05 Sep 2004 [GPGPU /Audio and Signal Processing]

[peacenik]

thanks guys!... keep the info coming .

[Phaidon]

Wrong forum??
"Music Computers" maybe?