Not to state the obvious but wouldn't you think that Apple will keep this all secret as long as they can.
The obvious likely path is a custom designed Apple chip either A8 or A9 based.
I'd say very close. So close that they can influence the GPU design team to support OpenCL.
It's a waiting game as there have been no significant leaks or announcements.
Dav
I'd agree - IMG kept the licensing deal schtum for months regarding Samsung/Apple. The stock buyup was unforeseen.
Who knows how far back Apple has gone with Nvidia and Imagination Techonologies regarding OpenCL. High level acknowledgement, if not involvement from both:
Tony King-Smith, VP of Marketing at Imagination Technologies:
“Imagination is delighted to have been involved in the authoring of OpenCL, which we see as a significant development for the future of GP-GPU based computing for multimedia.”
Tony Tamasi, Senior VP of Technical Marketing at NVIDIA:
“OpenCL adds fuel to the most exciting parallel computational revolution of our generation – GPU Computing. It also provides another powerful way to harness the enormous processing capabilities of our CUDA-based GPUs on multiple platforms.”
Michael McCool, founder and Chief Scientist at RapidMind:
“As a provider of a high-level parallel programming platform, RapidMind is excited about the availability of a new standard for targeting compute devices through a single API. The low-level access to a variety of devices provided by OpenCL will allow our platform to expand to new devices more quickly than ever before.”
Just look at
Rapidmind and their
PR releases
A "development and runtime platform that enables single-threaded, manageable applications that fully leverage multi-core processors. With RapidMind, developers continue to write code in standard C++ and use their existing skills, tools and processes and the RapidMind platform then “parallelizes” across multiple cores."
Will be intersting to follow their
blog, and see if they're around at at WWDC, if not talking more prior to SIGGRAPH in August.
With
OpenCL specification announced, then 6 months later validated to 1.0, they've already got some drivers out, and will have had another 6 months since December by WWDC.W e've only seen OpenCL in relation so fate to multi-core CPUs and GPUs, but they also mention "Cell-type architectures and other parallel processors such as DSPs."
HPC:
If they succeed, that's got to be some kind of industry spec development record -- basically from prototype to final in 6 months. I think the IEEE study group that was working on the 40/100 Gbps Ethernet standards took that long just to decide on the seating arrangement.
Would be interesting to go through the list of those
Khronos members announced as participating with OpenCL ( .e.g 3DLABS, AMD, Apple, ARM, Barco, Broadcom, Codeplay, Electronic Arts, Ericsson, Freescale, HI, IBM, Intel, Imagination Technologies, Kestrel Institute, Motorola, Movidia, Nokia, NVIDIA, QNX, RapidMind, Samsung, Seaweed, Takumi, TI & Umeå University.) and see what angles these companies are coming from.
e.g.
- ARM - Provider of embedded RISC microprocessors, peripherals and system-on-chip (SoC) designs
- Freescale Semiconductor - Embedded processors (microcontrollers, DSPs, Comms processors); currently focused on providing products to the automotive, networking and wireless communications industries.
- Imagination Technologies Group - develops, licenses and supplies market-leading 2D/3D graphics, DV, DSP, audio & speech technologies Intel - Larrabee? The PR stuff highlights OpenMAX - API to help comprehensive streaming media ("The OpenMAX API will be shipped with processors to enable library and codec implementers to rapidly and effectively make use of the full acceleration potential of new silicon - regardless of the underlying hardware architecture.") (Has anyone really heard about OpenMAX before? Just wondering).
- Samsung Electronics - the world's largest producer of color monitors, color TVs, memory chips, TFT-LCDs and VCRs.
- will use OpenGL ES and OpenMAX to implement solution involving graphics and multimedia applications for its various products and solutions.
- Texas Instruments - provides innovative DSP and analog technologies, and also Sensors & Controls, and Educational & Productivity Solutions.
Nvidia demo'd the
OpenMAX IL 1.0 standard back in
2006 - "The OpenMAX IL 1.0 (Integration Layer) specification defines media component interfaces to enable the rapid integration of accelerated codecs into streaming media frameworks on embedded devices. OpenMAX IL is the first of three layers of the overall OpenMAX standard that will provide comprehensive streaming media codec and application portability by enabling accelerated multimedia components to be developed, integrated, and programmed across multiple operating systems and silicon platforms. "
"OpenMAX IL will enable NVIDIA to put the full power of our advanced video, audio, and imaging silicon into the hands of software developers via a non-proprietary, open standard," said Michael Rayfield, general manager of the handheld GPU business unit at NVIDIA. "NVIDIA is a major contributor to the development of OpenMAX, and we are committed to provide leading-edge handheld media acceleration solutions through open standards."
Is this something Apple would get into? (The DL part at least?) -
OpenMAX is supported by TI in it's OMAP and OMAP Vox platforms. Nvidia supports, there are OpenMAX DSP functions for ARM11 and Cortex-8. And it's turning up more recently for Symbian - article
here
I wonder if 3.0 is based on SL or on Leopard?
That would be interesting if 3.0 is the same sort of progress as SL is for the Mac.
Kristof Beets was talking OpenCL at the April conference. OpenCL is SL.
video editing ?
SL's Quicktime X
MobileMe/youtube video sharing? SL
Multicore POWER VR? Points to SL usage
Microsoft Exchange Support?
DSP & OpenCL
Regarding DSP (Digital Signal Processing) and OpenCl - anyone have any ideas where that would go? I've seen the Register mentions you could use it to handle specialized tasks for which it was created - audio processing, for example. (XMOS link?) 7353143
As a tidbit, the same article says "After the tutorial, Nvidia's Trevett told The Reg that Intel's Larrabee crew - the engineers creating that GPU/CPU hybrid - are deeply involved in the evolving OpenCL spec."
Why would Nvidia, with the seemingly
bad blood happening regarding the lawsuit, drop this in? Would Intel backdown on GPU, and let Nvidia do their thing within an Intel chip? 2011 will be fun! It's big for Intel - "the tera-scale research program has been the single largest investment in Intel's technology research and has partnered with more than 400 universities, DARPA and companies such as Microsoft and HP to move the industry in this direction." So it's not just a pintsized Pentium processor...
We're beginning to wonder if these guys aren't just passing disses while sharing a cold one afterwards just to get attention, but being that it's more fun to envision suits from rival firms intensely angry with one another, we'll just keep believing this actually isn't a joke.
And potentially profitable - Nvidia's got bounties on new software -
Their
GPU Ventures Program giving $0.5m to $5m (£0.36 to £3.6m) a project...
