LONDON – Vivante Corp., a supplier of graphics processing unit (GPU) IP cores, has said that its 2-D GPU display and pixel composition engine has been now adopted by 5 of the top 10 application processor vendors and used in over 30 SoC designs.
The composition and GUI engine is a 2-D and pixel processing pipeline that supports up to 8K by 8K resolution, the company said. It also argues that running a 2-D graphics block for those things that need it, such as the user inteface and font rendering, while keeping the 3-D GPU switched off and only used when necessary is a good power-saving strategy. This approach is already used in Qualcomm's Snapdragon processors.
Vivante (Sunnyvale, Calif.) did not name it's licensees. It had previously claimed a total of 40 licensees using its GPU IP cores.
Vivants said its technology can blend, scale, animate, layer, filter, process, and combine multiple input sources (3D, 2D, HD video) to compose for final viewing. The IP block can be combined with Vivante's complete line of ultra-low power, high performance 3-D engine or other 3-D cores The block supports all CPU architectures; including ARM, MIPS, x86 and PowerPC, Vivante said.
Vivante CEO Wei-Jin Dai said, "Our vision is
this innovative technology should be in every application processor that powers
high resolution Android screens. We have succeeded in developing an important
piece of the visual ecosystem as market demand for better graphics, real time
effects, and instant UI response keeps growing. With SoC designers at a cross
road of balancing visual performance and power, our IP will help bridge this gap
by being a fundamental piece of the solution and lowering the overall system
power. Consumers can have a memorable user experience and solutions providers
can create the next wave of popular multimedia devices with GUI acceleration
that steps it up to the next level."
Not only does the approach make sense, but as I commented in the article it is already being done as a matter of course in many SoCs.
Vivante's real advantage is probably not that their IP is essential, but that it is good enough and a lot lower cost than similar from ARM or Imagination.
Cool idea. Sounds like what would be nice would be a unified instruction set and load balancer to keep all the ALU and graphics units busy. Now where have a heard of that before?
It looks like Vivante has some key IP and customers. Since they are small then they might get snatched up. It looks like Qualcomm made a comment above and I just saw this about TI's 4470 Apps processor: http://www.anandtech.com/show/4413/ti-announces-omap-4470-and-specs-powervr-sgx544-18-ghz-dual-core-cortexa9
2D doesn't have the flash and crowd appeal as does 3D, but I'd guess that the vast majority of computer applications do not need 3D rendering. Given that, it makes sense to off-load 2D to a different processor.
What would be nice is if, while running 2D, the GPU could be made available for high intensity computation. That wouldn't help with power consumption, except when the unit was both not running 3D and not doing the complex math, but it could give greater capabilities to otherwise less powerful systems.
Agreed, but the winners and lossers are not easy to predict in the early stages of a rapidly expanding market. I recall there were expectations the upstarts challenging 3dfx in the 1990s could not survive. We all know how that one turned out.
I see a similar approach on the multicore CPU side with the inclusion of smaller CPU cores for specialized tasks in the recent application processor product announcements. This type of specialized GPU seems like a natural extension of this thinking on the part of the AP vendors to graphics tasks. I can imagine the latest generation of AP GPUs must account for a significant battery drain during typical user interaction if they are indeed activated for blending and BLTing things to the screen. A right sized stand alone processor for this problem seems only natural and applies some of the lessons learned the early days of video cards.
That approach really makes a lot of sense. I assume the 3D engine is pretty large compared to the 2D engine, so if you have the smaller engine (less transistors, less die area) do more stuff, then there will definitely be major power savings. Plus, if I can use my phone or tablet for a few more hours I would definitely like that technology inside.
Join our online Radio Show on Friday 11th July starting at 2:00pm Eastern, when EETimes editor of all things fun and interesting, Max Maxfield, and embedded systems expert, Jack Ganssle, will debate as to just what is, and is not, and embedded system.