VR-Zone is claiming that this is the successor to NVIDIA's SHIELD portable gaming system. An AnTuTu benchmark was found for a device called, "NVIDIA test model(SHIELD)" with an "NVIDIA Gefroce(Kepler Graphics)" GPU, typos left as-is. My gut expects that it is valid, but I hesitate to vouch the rumor. If it even came from NVIDIA, which the improper spelling and capitalization of "GeForce" calls into question, it could easily be an internal prototype and maybe even incorrectly given the "SHIELD" (which is properly spelled and capitalized) label.
Image Credit: AnTuTu.com
As far as its camera listing, it would make sense for the SHIELD to get one at standard definition (0.3MP — probably 640×480). The fact that the original SHIELD shipped without any, at all, still confuses me. The low resolution sensor still does not make sense, seeming like an almost pointless upgrade, but it could be used by NVIDIA for a specific application or built-in purpose.
Or, it could be an irrelevant benchmark listing.
Either way, there are rumors floating around about a SHIELD 2 being announced at E3 in June. It is unlikely that NVIDIA will give up on the handheld any time soon. Whether that means new hardware, versus more software updates, is anyone's guess. The Tegra K1 would have been a good launching SoC for the SHIELD, however, with its full OpenGL 4.4 and compute support (the hardware supports up to OpenCL 1.2 although driver support will apparently be "based on customer needs". PDF – page 8).
Waiting. Seeing. You know the drill.
Please clarify, in one
Please clarify, in one paragraph you said both “As far as its camera listing, it would make sense for the SHIELD to get one at standard definition (0.3MP — probably 640×480).”
And then two sentences down “The low resolution sensor still does not make sense, seeming like an almost pointless upgrade…”
It sounds like you’re talking about the 640×480 camera in both instances, but you’re not sure if it “makes sense”?
Yeah, that was unclear. It
Yeah, that was unclear. It makes sense that the SHIELD would have got one off-the-bat but, since the first one didn't, it doesn't make sense as an upgrade. I cannot really imagine JHH going on stage and being like, "We listened to your feedback and we put a camera in! We recycled it from a 2005 Motorola flip phone."
Sorry for the confusion.
Hopefully it’s just a
Hopefully it’s just a placeholder camera then! I don’t know what the average phone has as a front-facing camera, but I think my Galaxy S4 has a 2MP camera. I wouldn’t imagine them going much higher than that, would you? … although, I never have really understood why they make the front-facing camera so much shittier than the outward one. : P
Yeah, the front facing camera
Yeah, the front facing camera is 2MP (1080p). It is a lot better than 480p (despite the iPhone 4S having one). Also, your rear-facing camera is 13MP.
As I said, NVIDIA could have added it for a specific application (if the benchmark listing is even valid). If so, great. If it is intended to be used with, for instance, Google-produced Android apps such as Google Translate, especially if it is rear-facing, 640×480 is quite low.
So those Denver cores are 7
So those Denver cores are 7 way superscalar, according to the whitepaper(PDF), are going to be competing with the other wide order custom ARMv8 ISA based superscalar design from Apple(A7 cyclone). It’s too bad that most of the Design wins of these K1 SKUs will be stuck running Android only, as they could run a full Linux distro on any tablet, the Apple A7 was compared to the Intel core i series by Anand, but the jury is still out on the Denver core.
The current 4 core SKU is still 32 bit ISA, but its GPU could make it great graphics tablet to compete with the more expensive graphics tablets on the market, and if Nvidia’s Denver core benches well against Apple’s Cyclone core, then their should be a market for a full Linux based graphics tablet, with better graphics than Intel’s moblie offerings, and lower to middle range core i series integrated GPUs.
It would not surprise me in the least if Apple does release an ipad pro, running OS X, but that will probably cost as much as a low end cintiq graphics tablet. Will the Apple’s rumored ipad pro, and phones get the PowerVR wizard GPU with hardware raytracing, and an Apple tablet running full OS X on an A8(?), would be more affordable than most of the professional graphics tablets on the market. The Nvidia K1’s desktop class GPU microarchitecture, and full desktop driver support is its best selling point, but the PowerVR wizard’s hardware raytracing will force Nvidia and AMD to come up with some hardware solutions for raytracing also! the Mobile market is the focus of much of the innovation and shows that where there is much competition there is much innovation.
That’s where Nvidia is going.
That’s where Nvidia is going. To laptops and then desktops. They just need the SoC to become more powerful and then SteamOS to start getting support and market share. After that we could see convertibles as android/SteamOS dual boot systems and latter even desktops.
If I haven’t completely misunderstood what NVlink is, just think of a desktop with an Nvidia motherboard housing an Nvidia SoC and having one or two NVlink slots for hi end Nvidia graphics cards running SteamOS.
From what I have been
From what I have been reading, Nvidia, and Samsung are part of IBM’s OpenPOWER Consortium/Fondation along with others, and IBM is going to license the Power CPU IP, in the same manner as ARM Holdings licenses it’s ARM IP, through a licensed IP Business model. AMD, Apple, as well as others may also be able to license the Power processor IP/ISA from IBM as well.
IBM is partnering with Nvidia to utilize Nvidia’s GPUs in IBM’s server/HPC/analytics products market. I also think that much of Samsung’s 14nm Fab technology was developed through IBM’s technology foundation partnerships, GlobalFoundries is also an IBM technology partner. IBM probably wants to secure plenty of third party foundry access, now that IBM wants to get out of the Chip fab business(except for some research Fabs). IBM is losing money on the hardware side.
NVlink was derived for IBM’s (Coherent Accelerator Processor Initiative), and IBM is seeding the market with IP, in which IBM hopes to benifit by having a constant supply of CPU/GPU parts to go in its power based server products, that run IBM’s proprietary OS. Everyone else will run power on linux, and IBM will have the best of both worlds, a plentyfull supply of power based parts from third party suppliers (Samsung, GlobalFoundrys), and Its real Money Maker, IBM’s proprietary OS/ecosystem, where IBM’s real profits are made. This all harkens back to IBM’s orginal intent in forcing Intel to cross license the x86 16/32 bit ISA with AMD, a cheep and competitive supply of x86 parts for the IBM PC, and now IBM wishes to do the same with its power based CPUs by opening up the power ISA/IP for license.
Gefroce? Really?
Gefroce? Really?
That’s some whitepaper(PDF)
That’s some whitepaper(PDF) it reads more lake a marketing brochure! Anyone know of any real whitepapers or datasheets out there that are not behind a must be a certified developer restriction. Please file shareing sites get a site dedicated to hardware data sheets, real whitepapers, and technical manuals. Real whitepapers use the proper computer sceince/software engineering terminology and not Nvidia’s marketing buzzwords and terminology! It’s no wonder Nvidia has so few Design wins with its Tegra line, or so few developers targeting the platform. Better get more info out there, or Bemma and Mullins may just take more market share.