Slashdot Mirror

'Furthermore the GPU design that replaces Imagination's designs will be, according to Imagination, "a separate, independent graphics design."'

Imagination does not acknowledge Apples claims, it actual fact Imagination says the exact opposite.

"Apple has not presented any evidence to substantiate its assertion that it will no longer require Imagination’s technology, without violating Imagination’s patents, intellectual property and confidential information"

Apple were also one time in talks to acquire Imagination's technology outright.

"From time to time, Apple talks with companies about potential acquisitions. We had some discussions with Imagination, but we do not plan to make an offer for the company at this time." March 2016

"Apple can't develop a new GPU in isolation -- that any GPU developed by the company would still infringe on some of Imagination's IP. "

Why is that the case? I don't see AMD, Intel, or nVidia among their licensees, and they make GPUs. Maybe they have a patent for "GPU, but on an Apple product."

And it looks like Imagination's first GPU (by the name PowerVR) came out in 1996. So it seems that the foundational patents would be expired by now.

single-thread performance is king for games

It was, some time last decade. Apparently you have not heard about Vulkan. Applications are going that way too, and, well, everything.

It's also unclear where, or even if, you can buy the thing (having CPU from Soviet Russia, or at least Putinstan, would be cool :-). So it seems like a different version of something like the Ci20 Creator, which has been around for awhile.

Here's a quick demo we've written for SIGGRAPH that compares Vulkan and OpenGL ES performance for Android on a prototype driver.

Great question (I get asked about this a lot). Here are a few points:

(1) Hardware multi-threading support: MIPS offers SMT support for the latest Warrior CPUs; for a slight increase in area (~10%), you can scale up the number of hardware threads (1, 2, 4) and get a 40-50% boost in performance. ARM CPUs do not support SMT and can only scale in number of cores.
(2) Better hardware virtualization support: MIPS Warrior CPUs support hardware virtualization at the low end (e.g. microcontrollers) as well as the high end (application processors). ARM CPUs support hardware virtualization only at the high end. Moreover, MIPS CPUs support multiple trusted execution environments (up to 7 in MCUs, up to 31 at the high end) while ARM CPUs have only one TEE.
(3) Better raw DSP performance: MIPS Warrior CPUs offer superior DSP performance vs. equivalent ARM CPUs (e.g. up to 2.5x better DSP performance in MCUs)
(4) Better performance efficiency: MIPS CPUs offer equivalent performance but at smaller area/power consumption over equivalent ARM cores (e.g. up to 30% area savings at the cluster level and 40% savings at the core level relative to similar performance competition)
(5) More mature 64-bit ecosystem in networking and embedded: MIPS 64-bit CPUs have a rich history in high-performance enterprise networking (examples include Broadcom XLP and Cavium OCTEON processors); there is a whole ecosystem formed around OpenWrt on MIPS for example.

Great question (I get asked about this a lot). Here are a few points:

(1) Hardware multi-threading support: MIPS offers SMT support for the latest Warrior CPUs; for a slight increase in area (~10%), you can scale up the number of hardware threads (1, 2, 4) and get a 40-50% boost in performance. ARM CPUs do not support SMT and can only scale in number of cores.
(2) Better hardware virtualization support: MIPS Warrior CPUs support hardware virtualization at the low end (e.g. microcontrollers) as well as the high end (application processors). ARM CPUs support hardware virtualization only at the high end. Moreover, MIPS CPUs support multiple trusted execution environments (up to 7 in MCUs, up to 31 at the high end) while ARM CPUs have only one TEE.
(3) Better raw DSP performance: MIPS Warrior CPUs offer superior DSP performance vs. equivalent ARM CPUs (e.g. up to 2.5x better DSP performance in MCUs)
(4) Better performance efficiency: MIPS CPUs offer equivalent performance but at smaller area/power consumption over equivalent ARM cores (e.g. up to 30% area savings at the cluster level and 40% savings at the core level relative to similar performance competition)
(5) More mature 64-bit ecosystem in networking and embedded: MIPS 64-bit CPUs have a rich history in high-performance enterprise networking (examples include Broadcom XLP and Cavium OCTEON processors); there is a whole ecosystem formed around OpenWrt on MIPS for example.

Great question (I get asked about this a lot). Here are a few points:

(1) Hardware multi-threading support: MIPS offers SMT support for the latest Warrior CPUs; for a slight increase in area (~10%), you can scale up the number of hardware threads (1, 2, 4) and get a 40-50% boost in performance. ARM CPUs do not support SMT and can only scale in number of cores.
(2) Better hardware virtualization support: MIPS Warrior CPUs support hardware virtualization at the low end (e.g. microcontrollers) as well as the high end (application processors). ARM CPUs support hardware virtualization only at the high end. Moreover, MIPS CPUs support multiple trusted execution environments (up to 7 in MCUs, up to 31 at the high end) while ARM CPUs have only one TEE.
(3) Better raw DSP performance: MIPS Warrior CPUs offer superior DSP performance vs. equivalent ARM CPUs (e.g. up to 2.5x better DSP performance in MCUs)
(4) Better performance efficiency: MIPS CPUs offer equivalent performance but at smaller area/power consumption over equivalent ARM cores (e.g. up to 30% area savings at the cluster level and 40% savings at the core level relative to similar performance competition)
(5) More mature 64-bit ecosystem in networking and embedded: MIPS 64-bit CPUs have a rich history in high-performance enterprise networking (examples include Broadcom XLP and Cavium OCTEON processors); there is a whole ecosystem formed around OpenWrt on MIPS for example.

Cool, thanks for the feedback. I will try to mention which architecture each CPU implements in the future so that it's clear for everyone. Regarding the ISA changes, let me explain further. For the cases you've mentioned we offer a software/hardware compatibility strategy which includes trap-and-emulate, trap-and-patch, and binary translation. To this end, although you are correct that some legacy instructions have been deprecated, binaries should still run because of this three pronged approach. Have a look at http://www.imgtec.com/mips/arc... and please use our forums for any questions since the guys there are quite good at replying.

It is not a lure, the terms of the agreement are clear: 1) you are not allowed to put into silicon and 2) you must consult with Imagination before you change it. Most universities teach courses based on MIPS CPUs so having access to the real RTL code was something they were very excited about - this is relevant in the quotes from many well-known professors. http://www.imgtec.com/news/det... If these highly respected professors thought this was a lure and unethical, they would not have signed up for the programme. The reality is we've consulted with tens of universities and this is what they wanted (i.e. full access to the RTL) so we were able to give it after many months of hard work from our engineering team.

Does this do tile-based rendering?

I presume you intended this as a joke, but from the OpenGL/Vulkan comparison table in the overview: "Matches architecture of modern platforms including mobile platforms with unified memory, tiled rendering"

On that point - I'm not all that sure exactly what it does to support tiling. The PowerVR blog says "A render pass consists of framebuffer state (other than actual render target addresses), and how render targets should be loaded in and out of the GPU at the start and end of each render. This structure is the key object that allows tiled architectures like PowerVR to run at extremely high efficiency." but that it's not really all that clear to me how this makes a difference.

Imagination Technologies (PowerVR) posted this today with more in-depth info:
http://blog.imgtec.com/powervr...

The purpose of Vulkan is apparently to be a low-level alternative to the high-level APIs OpenGL and OpenGL ES.
Game consoles such as the Playstation series have had both high-level and low-level graphics API:s for many years. Using the low-level API means that you can squeeze out more performance, perhaps at the expense of more developer time. The application takes over more duties, such as resource management etc.
If your app is a game, then your resource management and shaders are often pretty much static anyway.

This is probably not a useful answer because I suspect you want a solution that actually exists, but: Imagination is quite strongly pushing their new [PDF] ray tracing technology, which is essentially a hardware accelerator for ray/AABB and ray/triangle intersection tests, tightly coupled with a normal PowerVR GPU. They claim it's hugely more efficient than emulating the same behaviour with GPGPU. I guess the downsides are that if you don't care about power, you can easily buy a desktop GPU and brute-force it with GPGPU; and that the old accelerator cards seem pretty expensive, and the new SoC-based solution doesn't exist as an actual product yet. But it might become an interesting solution in the near future.

This is probably not a useful answer because I suspect you want a solution that actually exists, but: Imagination is quite strongly pushing their new [PDF] ray tracing technology, which is essentially a hardware accelerator for ray/AABB and ray/triangle intersection tests, tightly coupled with a normal PowerVR GPU. They claim it's hugely more efficient than emulating the same behaviour with GPGPU. I guess the downsides are that if you don't care about power, you can easily buy a desktop GPU and brute-force it with GPGPU; and that the old accelerator cards seem pretty expensive, and the new SoC-based solution doesn't exist as an actual product yet. But it might become an interesting solution in the near future.

Heh...if it were trash, you'd not see it being one of the primary GPUs for mobile devices.

The problem isn't PowerVR itself. It's proprietary drivers on their stuff not being available that's always been the problem. Else you'd probably be singing their praises like most do NVidia for the problem space.

Mods, lay down the damn crack pipe...this one's not "informative", nor is it even accurate. Proof:

Rockchip RX3168 uses SGX GPU
Apple uses SGX in their iPads...
Ingenic uses it with their MIPS Android SOC
MediaTek uses it

And the list goes on and on. Unless this one uses AMD (well and now NVidia with Kepler...) or can convince Broadcom to make their top-end Videocore IV stuff available to use, they're talking out their backside out of box. It's crap because you can't easily get drivers for it and it's closed- and the only difference between them and NVidia on X86 is that NVidia makes closed drivers available, ImgTec doesn't. The same can be said for Adreno, Mali, and a few others. Now, if ImgTech could ass themselves to make available their drivers for X86 Atom boards, things would be better there.

I would suggest that the much bigger problem with a lack of price means that it's currently impossible to get one at all, since the giveaway period has also expired.

but then there is the fact that most RISC's use same registers for both FPU and integer

With the minor exceptions of Alpha, PA-RISC 1.x and 2.0, POWER/PowerPC/Power ISA, MIPS, and SPARC.

The hit isn't a very big one:

"with the hardware offload the battery lasted up to 36% longer"

http://blog.webmproject.org/20...

And with each faster processor generation, the difference gets smaller and smaller still.

Followed the link but couldn't see where it showed actual power consumption of the hardware decoder they used (their own I guess?) but given that an ARM CPU might consume around 500mW whereas an H.264 hardware decoder doing HD uses 10mW, either the screen is using a huge amount of power or their hardware leaves a bit to be desired.

Imagination Technologies, owners of PowerVR, recently became members of the Open Handset Alliance four months ago. Open Handset Alliance is the Android booster org. Before this they were a strictly proprietary driver company, and Android devices that used their tech had binary blobs. The binary blobs aren't gone yet, but they soon will be replaced with open source licensed drivers and actual hardware specifications.

So Microsoft needs there to be a mobile GPU tech company that has secret drivers to sell their mobile software on platforms that can't be made useful with a software flash. They cast about and set their sights on nVidia, who has already signed their devil's deal to keep how their PC hardware works a trade secret. They probably promised nVidia something useless to get this - that's their usual course. Now Microsoft's puppet hardware ODMs will build Microsoft nVidia GPU-based tablet platforms that can't run Android, won't sell, and have to be dumped all over the place like Surface RT is now. Expect Surface RT 2, whatever it's properly named, to use this tech. In the end nVidia gets hosed - again. If you sup with the devil, use a long spoon.

Intel used Imagination Tech in their Atom line as well, and that's why you can't get good Linux drivers for those otherwise sweet mini-itx boards. Yet. They're coming. Intel has dropped them though for some reason now in favor of in-house tech.

It's really hard to track the machinations in GPUs.

Don't get too excited. Some intel chips use PowerVR, which has no OSS driver (from intel as well as from anyone else) see here

The Tegra 3 chip that's showing up in phones this spring and Transformer Prime tablet now is about 7.2 GFLOPs. That's more than enough to be top 10 in 1993. Current ARM architectures might go all the way up to fast enough to take that number one spot in reference sample designs now but they consume too much power to go in your pocket on retail shelves as yet. Maybe in a year or two.

Mali T658 and PowerVR are two to watch here. Mali is supposed to go up to 350 GFLOPs. It still amazes me that in 1993 that machine cost about $70 million in today's money and you can almost match it today for under $500.

You're thinking of it like there's dedicated h264 hardware in, say, the iPhone. There isn't. There's hardware that accelerates decoding of h264... that same exact hardware can be used to decode VP8.

I don't think that is true. The iPhone 3GS + 4 onwards have some form of Imagination Technologies VXD video decoder hardware and presumably those only support some subset of the following " H.264, VC-1, MPEG-4, MPEG-2, MPEG-1, JPEG, AVS, Sorenson Spark, Real and On2 VP6".

Hmm, perhaps the iPad is said to only display 720p video because that's how big its display is???

That is my theory as well. The iPhone 3GS is rumored to have a VXD video decoder and that hardware supposed to handle 1080p (Blu-ray). It would seem strange if the iPad had weaker hardware.

Or maybe advertising your use of Linux moves you from being a mere annoyance to a mortal enemy of the Redmond cabal.

:-) ... but I think a far more likely limitation would be that there is probably no version of Windows CE for the processor in the radio, i.e. Imgtec/Pure's Meta

Jeez, could we get a few more acronyms and buzzwords in this summary please?

Sorry for yet another "acronym" but if they want to be running an HD media player, my guess it's not so much the SGX technology they want to use, but the vxd (more info here) unit that is also in the chipset.

I'm using my own N95 http://www.nseries.com/n95/ as:

• an IRC client: http://mirggi.net/ (native Symbian software)
• a SSH client: http://s2putty.sourceforge.net/ (Putty is ported to and runs on Symbian natively)
• a podcast player: Nokia Podcasting (http://blogs.s60.com/nokiapodcasting/). I can download new episodes on the fly and listen to them when I want. I don't need a computer to download the episodes. MPEG-4/H.264 video podcasts work too.
• an Internet radio (Shoutcast) client: http://opensource.nokia.com/projects/s60internetra dio/index.html All the Internet radio stations just when I want. Open Source.
• an FM radio and MP3/AAC player. Any headphones with a 3,5mm plug work fine.
• a 5 megapixel digital camera: http://www.flickr.com/cameras/nokia/n95/
• a 640x480 30fps MPEG-4/AAC video camera.
• a modem for my laptop. Thanks to HSDPA (http://en.wikipedia.org/wiki/Hsdpa) I get about 120kB/s downstream and 44kB/s upstream (yes, kilobytes) with the current 1,8Mbps HSDPA network. 90ms pings. I have an unlimited packet data contract from my mobile operator. The operator is currently software upgrading the base stations to support 3,6Mbps HSDPA, which doubles the downstream speeds.
• a SIP VoIP client: The SIP standard is supported by the device natively. And the Internet call functionality is well integrated to the user interface. I can use the normal phone book to call via the Internet. Instead of normal voice or video call, I just select Internet call from the menu. Internet calls work over 802.11g 54Mbps WLAN at home, and over the mobile packet data network on the go, thanks to HSDPA. Works great with Gizmo (http://www.gizmoproject.com/), for example.
• a web browser and RSS feed reader: RSS feeds are supported by the excellent S60 web browser, which is based on Apple Webcore/KHTML: http://opensource.nokia.com/projects/S60browser/. It's a full featured web browser and not a toy. Web sites can be zoomed in/out to fit the screen. Opera can be installed on the device too.
• a gaming device: I'm not a enthusiastic gamer though. I'm just playing the preinstalled games. btw. N95 has got hardware accelerated OpenGL by PowerVR: http://www.imgtec.com/PowerVR/Products/Graphics/MB X/index.asp
• a Push-email client. I receive email as soon as it is available on the IMAP server. I can open ZIP attachments, Word/Excel documents, PDF files and view them on the device.
• a GPS and a map: N95 has got an integrated GPS receiver. Nokia Maps software is preinstalled on the device. Free detailed maps for over 150 countries are available. And Google Maps works on it too: http://www.google.com/gmm/. I also use N95 to track my work-outs with Nokia Sports Tracker http://research.nokia.com/research/projects/Sports Tracker/.

I can install any S60 3rd edition and Java ME software on this phone. No restrictions. Also the SDK is available freely: http://forum.nokia.com/ Symbian is also fast, and supports multitasking of applications natively. I can program it even with Python (http://wiki.opensource.nokia.com/projects/PyS60) if I want.

Hot-swappable 2GB Micro SD cards work as stora

As mobile device makers begin to put OpenGL ES chips into their devices, like this, http://www.imgtec.com/PowerVR/Products/Graphics/MB X/index.asp?Page=2 and if these devices are used as X servers (i.e., remote display terminals) then the only thing that remains for fantastic user experiences is people to write X apps to make the devices specifically useful in the different scenarios in which all people find themselves day in and day out.

1. "We will reshuffle development organizations to be more efficient and to make only games that will sell well," Sega President Hisao Oguchi said.

This quote from news.com makes me worried about exactly what kind of games they're really talking about. I do hope we don't see Sammy churn out video games that are as generic as pachinko machines. Same thing over and over with a different graphical template. I shudder to imagine.

2. Part of Sega's allure for Satomi--who owns 27.6 percent of Sammy's shares outstanding--is using Sega's titles in Sammy's "Atomiswave" game machines for arcade use. Sega's arcades will also likely use the Atomiswave machines.

What does THIS mean for Sega's plans for a new arcade based off PowerVR's brand new Series 5 hardware? While probably similar to Atomiswave it's certainly not the same thing.

Whether Sega produces more games for the console or not really doesn't concern me as much as whether I'd ever want any of these generic quarter munchers to begin with.

Ahh.. good eye!

They appear to have some PC cards

They also have drivers for Windows XP and Linux. Surprisingly the linux drivers seem pretty robust.

Unfortunately the latest cards are only based off the KRYO 2 SE or PowerVR Series 3 chipset which is a bit dusty considering Series 2 is what they used in the Dreamcast and PowerVR Series 5 chips are the ones they're basing the new SEGA arcade hardware on. I wonder how a Series 5 board would fare against Nvidia and ATi's offerings...

Ahh.. good eye!

They appear to have some PC cards

They also have drivers for Windows XP and Linux. Surprisingly the linux drivers seem pretty robust.

Unfortunately the latest cards are only based off the KRYO 2 SE or PowerVR Series 3 chipset which is a bit dusty considering Series 2 is what they used in the Dreamcast and PowerVR Series 5 chips are the ones they're basing the new SEGA arcade hardware on. I wonder how a Series 5 board would fare against Nvidia and ATi's offerings...

Actually, Videologic (I believe they were independant at this time. NEC ... bought them outright)

Not so. NEC make the chips. Videologic is a division of Imagination Technologies, indeed they weren't taken over, they just changed their name. It's all on their site. Oh, they're a UK company too (yay!)