Gráfica nVidia Pascal (Mobile, TITAN Xp, GTX 1000 series)

Pelo q percebi.... isto vai ser para tirar mais partido dos cpu... pq actualmente... a sobrecarca esta toda no gpu em aplicacoes. o cpu assim tb leva com a carga do gpu... torna as gpu mais eficazes energeticamente. E fazse uso dos cores do cpu. Em qualquet situacao no uso do gpu. Libertando os jogos de terem q se adaptar por software ao cores cpu. Basicamente isto ate boa noticia para os amd... q tem forca bruta a nivel de cores.
 
Nvidia CEO Talks About Pascal PK100 and PK104 GPUs – Will Be Produced with FinFETs and Feature HBM2
Hardware 7 hours ago by Usman Pirzada
Nvidia recently held their earnings call (via KitGuru) and the CEO Jen Hsun talked a bit about the upcoming Pascal architecture. As expected however, he didn’t reveal any new detail except rehash what is already known. He did hint that further details about the product will be “coming soon” and he can’t wait to talk about them. Pascal generation of GPUs will mark an important shift in Nvidia GPUs, shifting from GDDR5 memory to HBM2 – resulting in a huge leap in bandwidth.

A slide from Nvidia’s press deck showing Pascal being roughly 10x faster than Maxwell. @Nvidia Public

Nvidia could reveal further information on Pascal GPUs sooner than expected – CEO Jen Hsun can’t wait to talk about them
Pascal GPUs are slated to be pretty damn amazing. They will feature 4X the mixed precision performance, 2X the performance per watt, 2.7X memory capacity & 3X the bandwidth of Maxwell. Considering the fact that Nvidia’s color compression technology is already phenomenal, I am excited to see what a Pascal GPU with the same tech and massive bandwidth can offer. The PCI-E bottleneck remains as always – but that is quickly becoming irrelevant as more of the work is handled on-gpu thanks to the autonomy granted to the Graphic Processor by low level APIs such as DiretX 12.

“I cannot wait to tell you about the products that we have in the pipeline,” said Jen-Hsun Huang, chief executive officer of Nvidia, at the company’s quarterly conference call with investors and financial analysts. “There are more engineers at Nvidia building the future of GPUs than just about anywhere else in the world. We are singularly focused on visual computing, as you guys know.”

There are currently two Pascal cores that we know about – the PK104 and the PK100. The PK104 will be the GM204 successor while the PK100 will be the flagship Pascal GPU from Nvidia. Both will be manufactured on either the 14nm or 16nm node using FinFETs. Nvidia has been on the same 28nm node for around 3 years now but has managed to always deliver in terms of performance jumps per generation. Ofcourse, that will becoming exponentially difficult to maintain unless they shift to a lower process – something they plan to do with their next generation of GPUs. The use of HBM2 will let Nvidia put upto 32GB of video buffer on their GPUs and upto 1TB/s of bandwidth.

“We have found over the years to be able to focus on just one thing, which is visual computing, and be able to leverage that one thing across PC, cloud, and mobile, and be able to address four very large markets with that one thing: gaming, enterprise, cloud, and automotive,” said Mr. Huang. “We can do this one thing and now be able to enjoy all and deliver the capabilities to the market in all three major computing platforms, and gain four vertical markets that are quite frankly very exciting.” …“I cannot wait to tell you all about it, you are just going to have to wait just a little longer,” – CEO, Nvidia

We were not able to spot any new details in the Earnings Call apart from what we already know, so that was basically a little recap on Pascal GPUs. Pascal Architecture is expected to succeed Maxwell Architecture and combined with DX12 should usher in a new era of Gaming prowess. With processors consistently and comfortably ahead of the demand curve, the onslaught of high resolution standards such as 4k have increased the pressure on better GPUs. Pascal could very well be poised to offer a lineup of one of the first 4K capable single-gpu cards at the 4k@60 fps standard.


Read more: http://wccftech.com/nvidia-ceo-talks-pascal-pk100-pk104-gpus-produced-finfets-hbm/#ixzz3ZinJQ6hv

http://wccftech.com/nvidia-ceo-talks-pascal-pk100-pk104-gpus-produced-finfets-hbm/
 
NVIDIA Tapes Out "Pascal" Based GP100 Silicon

Sources tell 3DCenter.org that NVIDIA has successfully taped out its next big silicon based on its upcoming "Pascal" GPU architecture, codenamed GP100. A successor to GM200, this chip will be the precursor to several others based on this architecture. A tape-out means that the company has successfully made a tiny quantity of working prototypes for internal testing and further development. It's usually seen as a major milestone in a product development cycle.

With "Pascal," NVIDIA will pole-vault HBM1, which is making its debut with AMD's "Fiji" silicon; and jump straight to HBM2, which will allow SKU designers to cram up to 32 GB of video memory. 3DCenter.org speculates that GP100 could feature anywhere between 4,500 to 6,000 CUDA cores. The chip will be built on TSMC's upcoming 16 nanometer silicon fab process, which will finally hit the road by 2016. The GP100, and its companion performance-segment silicon, the GP104 (successor to GM204), are expected to launch between Q2 and Q3, 2016.
 
Nvidia Pascal GP100 GPU Flagship Will Pack A 4096-bit Memory Bus And Four 8-Hi HBM2 Stacks
EXCLUSIVEHARDWARE 17 hours ago by Khalid Moammer
Nvidia’s big Pascal GPU code named GP100 will feature a massive 4096bit bus and four HBM2 stacks each up to 8-Hi. The upcoming Nvidia flagship Pascal chip set to debut on TSMC’s 16nm FinFET process later next year. We have confirmed with our sources that the GPU will be made with two different variations of stacked HBM2 solutions, however both will feature a massive 4096bit memory interface just like AMD’s flagship Fiji GPU launched last month.


The first variation will pack four HBM2 stacks, each will be 4-Hi and will be clocked at 1Ghz. This will go into the traditional consumer GeForce line of GP100 based products. The second variation is also equipped with four HBM2 stacks clocked at 1Ghz, however each will be 8-Hi.

In HBM stacking #-Hi denotes to the number of stacked DRAM dies, however this system does not take into account the additional base die which incorporates logic and memory PHY. So it’s only meant to specify how many DRAM dies are in the stack rather than the total number of chips in the stack. GP100 packages with 8-Hi HBM stacks will be limited to professional products, including Quadro and TESLA GPUs, where huge memory capacities are essential.

Nvidia’s Big Pascal Slated For Release Next Year
We’ve heard just recently that Nvidia’s flagship Pascal code named GP100 has actually taped out on TSMC’s 16nm FinFET manufacturing process last month. Interestingly we’ve also learned just three days ago that AMD has also taped out two FinFET chips last quarter. It’s absolutely not a coincidence that both companies completed their designs at exactly the same time. Both are aggressively pushing for a debut of their next generation FinFET based GPUs next year.

TSMC’s new 16nm FinFET process promises to be significantly more power efficient than planar 28nm. It also promises to be a bring a considerable improvement to transistor density. Which would enable Nvidia to build faster, significantly more complex and more power efficient GPUs.

TSMC’s 16FF+ (FinFET Plus) technology can provide above 65 percent higher speed, around 2 times the density, or 70 percent less power than its 28HPM technology. Comparing with 20SoC technology, 16FF+ provides extra 40% higher speed and 60% power saving. By leveraging the experience of 20SoC technology, TSMC 16FF+ shares the same metal backend process in order to quickly improve yield and demonstrate process maturity for time-to-market value.

Although notably, unlike Nvidia which has confirmed that its Pascal GPUs will be manufactured using TSMC’s 16nm FinFET process, AMD has yet to announced whether the Arctic Islands family of GPUswill be made on TSMC’s 16nm or Samsung’s 14nm process. Both nodes are very similar, so which process AMD ends up using will be primarily dictated by yields and time-to-market.

Also unlike Nvidia, AMD has a much more powerful incentive to launch its next generation of FinFET GPUs first. This is because the company has priority to HBM2 capacity – which is going to be limited initially – as a result of co-inventing the technology with Hynix. By pushing its graphics products to launch first AMD can establish two competitive advantages over its rival. The first obvious advantage is being first to market by launching its products earlier than its rival. But most importantly this enables AMD to capture much of that initial HBM2 capacity away from Nvidia and extend its time-to-market lead substantially. This could create an interesting market dynamic but whether it can succeed remains to be seen.

Obviously Nvidia realizes that this play is in the cards, no pun intended, and will undoubtedly bide its time wisely honing its chips. GP100 has already been taped out so there’s not much that can be done to the chip’s floorplan, however Nvidia can still use that extra time on post-silicon work. Nvidia can also spend more time working on its smaller Pascal chips which haven’t been taped out as of yet.

Apart from HBM2 and 16nm there is one big compute-centric feature that Nvidia will debut with Pascal. And it’s NVLink. Pascal will be the first GPU from the company to support this new proprietary server interconnect.


The technology is aimed at GPU accelerated servers where the cross-chip communication is extremely bandwidth limited and a major system bottleneck. Nvidia states that NV-Link will be up to 5 to 12 times faster than traditional PCIE 3.0 making it a major step forward in platform atomics. Earlier this year Nvidia announced that IBM will be integrating this new interconnect into its upcoming ***** server CPUs.



NVIDIA® NVLink™ is a high-bandwidth, energy-efficient interconnect that enables ultra-fast communication between the CPU and GPU, and between GPUs. The technology allows data sharing at rates 5 to 12 times faster than the traditional PCIe Gen3 interconnect, resulting in dramatic speed-ups in application performance and creating a new breed of high-density, flexible servers for accelerated computing.
Unlike Maxwell, Nvidia has laid major focus on compute and GPGPU acceleration with Pascal. The slew of features and new technologies that Nvidia will debut with Pascal emphasize this focus. Including the use of next generation stacked High Bandwidth Memory, high-speed NVLink GPU interconnect and support of mixed precision for the acceleration of mobile applications to push on mobile perf/watt. We can’t wait to see Pascal in action next year, but until then stay tuned for the latest.



Read more:http://wccftech.com/nvidias-gp100-p...-bit-memory-bus-8hi-hbm-stacks/#ixzz3gcJ9Nr3K

http://wccftech.com/nvidias-gp100-pascal-flagship-pack-4096-bit-memory-bus-8hi-hbm-stacks/
 
Pascal com 17 mil milhões de transístores.

Pascal technology will have as many as 17 billion transistors under the bonnet, Fuzilla can exclusively reveal.

Pascal is the successor to the Maxwell Titan X GM200 and we have been tipped off by some reliable sources that it will have more than a double the number of transisters. The huge increase comes from Pascal's 16 nm FinFET process and its transistor size is close to two times smaller.

Nvidia and AMD are making their GPUs at TSMC and the Taiwanese foundry has announced 16nm FinFET production runs. Intel and Samsung/ GlobalFoundries call their process 14nm. Our sources told us the branding depends which side of the transistor you look at, the longer or shorter. The size of the gate is almost identical for both 16nm and 14nm process.

Pascal has 17 billion transistors and it will be significantly smaller silicon than the Maxwell 28nm based GM200.

Nvidia will use second generation HBM for its Pascal GPU to get to a 32GB on the highest end card, This is 2.7 times more than the already impressive 12GB used on Titan X. The second generation HBM or HBM 2.0 will enable 8Gb per DRAM die, 2Gbps speed per pin and 256 GB per second Bandwidth/ stack.

The first generation offers 2Gb Density per DRAM die, 1Gbps speed per pin, 128 GB/s GB per second Bandwidth and maximum of 4 Hi stack chips with 4GB per HBM card. You saw this with Fiji cards.

HBM2 enables cards with 4 HBM 2.0 cards with 4GB per chip, or four HBM 2.0 cards with 8GB per chips results with 16GB and 32GB respectively. Pascal has power to do both, depending on the SKU.

The GPU looks great but it is coming in 2016, and not before. After Pascal comes the Volta GPU, but that will take a few years.

http://www.fudzilla.com/news/graphics/38304-nvidia-pascal-gpu-has-17-billion-transistors
 
Nó com problemas, vão já fazer um chip na capacidade máxima do nó, quando um site começado por FUD é tomado como fonte credível de noticias..
 
Nó com problemas, vão já fazer um chip na capacidade máxima do nó, quando um site começado por FUD é tomado como fonte credível de noticias..

Eles não dizem que o primeiro Pascal é a versão com 17 mil milhões de transístores. E a 14/16 nm, fica mais pequeno que o actual GM200
 
Sim, fica mais pequeno, mas em 16nm não tens aperture para 600mm^2, é mais na casa dos 300mm^2 de área máxima para esse nó.
 
Quando mudas de nó(tamanho de features, metalizações e afins) o tamanho máximo da apertura(máscara que é usada para fazer a exposição) tem um tamanho máximo menor.
 
a 980ti tem 8 bilhoes é assim tao pouco provavel eles duplicarem o numero de transistors? com a mudança para 14/16nm a area por transistor nao diminui para ~1/4?
 
E á quantos anos estão nos 28nm?
Não é á um ano, nem dois..

E não esquecer que os 14nm não são bem 14nm, os transistor são, mas as grelhas e metalizações não escalaram lá muito bem para 14nm e tens espaçamento maior, sim permite muito maior densidade, mas não é uma transição/redução perfeita.
 
Concordo com o senso, está noticia dos 17 mil milhões de transístores é treta. Apenas click bait de um site bem conhecido por publicar tudo e mais alguma coisa, sem confirmação nenhuma.

Mesmo após vários anos de 14nm, não deveremos ver um numero tão grande de transístores neste processo de fabrico, quanto mais logo no inicio.
Acredito mais que vejamos uns 10 mil milhões no inicio e uns 12 mil milhões no final.
 
Mas 14/16nm nao é o nó a seguir (seria 20nm, meio nó de 22nm) por isso nao é expectavel um salto maior? talvez nao o dobro mas perto disso, pelo menos depois de uns anos no processo?
 
20nm deu demasiados problemas, saltaram logo para os 14/16nm, mas estão sempre dependentes das foundries e de como a coisa corre, e não esquecer que de tape-out até produção ainda muita coisa tem de correr bem.
Eu não tenho nada contra a evolução de tecnologia, muito pelo contrário, simplesmente é sonhar um bocado demasiado alto(nunca se sabe se não me estou a enganar!) esperar um GPU daqui a meia duzia de meses com o dobro dos transistores.
 
Back
Topo