Nvidia is able to roll over in the mobile field?
At the opening of the CES 2014, nvidia released prototypes K1 a new generation of mobile processors, interesting is that nvidia with a model released two processors.192 Kepler GPU core architecture, make prototypes K1 in 3.0 the performance close to apple A7 GFXBench three times.Performance is very impressive.
The performance of the super
Nvidia in the mobile market has been doomed, this time with super performance prototypes K1 can turn over?Let's do a reading.
One, the truth of the prototypes K1
Actually according to the nvidia product roadmap, 2014 CES out should be Tegra5, which is the first prototypes K1, four ARM core processor A15 plus 192 Kepler GPU core architecture.
Four versions of K1
By using two 64 core in the second paragraph of the prototypes of K1, is actually prototypes 6, namely nvidia years ago became Denver, Denver is nvidia ready to challenge the Intel desktop high-performance ARM core processor, USES only the ARM instruction set, and not authorized by the ARM core, similar to apple A7 processor, the pursuit of high performance.
Because apple A7 was released in 2013, the mobile processors entered the age of 64, ahead of samsung, qualcomm also followed suit, so nvidia also had to release ahead of time, the core of Denver in prototypes in the second paragraph of K1.
Dual-core 64 K1
Denver engineering published in 2011
So, prototypes of K1 and millet 3 are now used in the first paragraph Tegar4 difference is not big, the difference only on the GPU.And in the second half of the second paragraph prototypes K1, ability is worth looking forward to the new product performance.
Second, the nvidia's ambitions
Prototypes K1 is focused on Kepler GPU core architecture, we can see it as a high performance 3 d processors, let us run the game more fun.
But more than that actually, nvidia GPU Kepler architecture not only can be used in 3 d rendering, more general, can be used to have very strong performance.
General-purpose computing can use less of the transistor, lower power consumption and fever to get higher performance than general CPU.
Nvidia prototypes K1 first realized the function of general-purpose computing on the mobile phone processor, that is to say, if the appropriate programming and task, prototypes K1 can explode out of 10 times the competitor's high performance, applied in the field, opponents cannot be applied to better user experience and more USES.Have absolute competitive advantage, this is the nvidia's ambition.
Third, nvidia's soft rib
Nvidia in the mobile space early, early in the iPhone before the advent of the era of WM, issued the mobile GPU nvidia -- Goforce series.
Nvidia Goforce a pioneer in mobile processors
Nvidia was subsequently found, however, are the need to integrate mobile field, independent GPU no market, the authorization of the nvidia then buy the ARM, the CPU and GPU integrated, that is the series of prototypes.Tegra2 is the short-term success, became the first dual-core mobile processors.
Enjoyed the prototypes of 2
But the core of the smart phone is not application processor (AP), but a baseband processor (BB), a mobile phone to communication first, and then the entertainment, bought the ARM authorization, the fusion of CPU and GPU is not enough, and fusion of BB.Otherwise, as long as qualcomm and MTK hands slightly on the supply, the success of the nvidia will like on the beaches of building collapsed.So nvidia acquired Icera baseband manufacturers, catch up on their own.
Unfortunately, in May 2011, since the acquisition, nvidia didn't digest complete launch can use integrated processor, this is the soft rib, nvidia prototypes 3 and 4 prototypes and stumbled are correlated with the soft rib.
Four, nvidia can move prototypes K1 roll over?
Through the analysis of the above, we can see that although prototypes K1 performance is strong, but the soft rib of the baseband did not resolve.Prototypes to do K1 mobile products, must match a pricey baseband processor, to do responsible for debugging of matching, while selecting qualcomm, MTK does not need to match, a lower cost.
, on the other hand, small mobile demand for high performance, and the demand is high, the power consumption of nvidia and no advantage.Nvidia phone general idea is very good, but need complex programming and suitable application scenarios, mobile phone is much more difficult than the PC (rendering video on mobile phones, dig? Distributed computing?)Does not reflect general advantage, a short time.
Advantage is difficult to reflect, soft rib haven't make up for, so the prospect of prototypes K1 is not wonderful, nvidia to turn remains to be later products.