Mal davon abgesehen, dass der dicke Pascal sowieso vermutlich auch nicht mehr dieses Jahr für Endkunden zu kaufen sein wird.
Hmmm, da liegen aber viele AMD Fans falsch mit dieser Vermutung,
Nvidia hält seine GPU HPC Roadmap & neuen CUDA 8 / 9 SDK Releases termingerecht ein (siehe Bild im Anhang) - nicht so wie AMD und deren FirePro Karten samt unzureichende OpenCL Treiber Unterstützung & über 1 Jahr verspäteter unvollständiger SDK ( Do-it-Yourself Support / OpenGPU Ecosystem & möchtegern copy&paste peudo AMD Cuda 'Boltzmann initiative' )
Cray CS-Storm and NVIDIA GPU Supercomputing Solutions | Cray Inc. /
Oak Ridge Leadership Computing Facility - Summit HPC 2017 with (Multiple) NVIDIA Volta GPUs per Node = 300 Petaflops /
Welcome to KAUST Supercomputing Laboratory | www.hpc.kaust.edu.sa
Nvidia wird die
HPC GP100 GPU ( 6.144 CUDA-Cores ) spätestens auf der Hauseigenen GTC (Nvidia - GPU Technology Conference) vom 4th - 7th April' 2016 @ San Jose Convention Center, in Silicon Valley
GPU Technology Conference 216 - #1 GPU Developer Event, samt den GP104 HighEnd / GP106 Gamer Pascal DX12+ GPU's vorstellen.
Dies wurde ja letztes Jahr, bereits so angekündigt! => Nvidia Pascal GPU is Just One Year Away - insideHPC / 23. März' 2015
">>> The new architecture will also include 3D memory, which will provide up to three times the bandwidth and nearly three times the frame buffer capacity of Maxwell GPUs. By reducing the memory bandwidth constraints, which limit the speed at which data can be delivered to the GPU; developers can build larger neural networks and accelerate the bandwidth-intensive portions of deep learning applications.
Pascal will have its memory chips stacked on top of each other, and placed adjacent to the GPU, much closer than previous architectures. This reduces the distance that bits need to travel from memory to GPU resulting in accelerated communication and better power efficiency.
This is in addition to up to 32 GB of memory — 2.7 times more than the newly launched Nvidia flagship, the GeForce GTX Titan X. One of the most anticipated features of the new Pascal architecture is NVLink – Nvidia’s high-speed interconnect, which links together two or more GPUs — which should increase the gains seen from the increase in available memory and bandwidth.
NVLink allows for double the number of GPUs in a system to work together in deep learning computations. In addition, CPUs and GPUs can connect in new ways to enable more flexibility and energy efficiency in server design compared to PCI-E. <<<"
Des weiteren warten schon
diverse Kunden (Grosskunden) im High Performance Computing & Forschungs-Bereich, auf die Auslieferung & den Einsatz jener neuen Nvidia PASCAL HPC GPU's ab dem 15. Mai 2016 - siehe
CCRT in France Acquires 1.4 Petaflop "Cobalt" Supercomputer from Bull - mid 2016 / insideHPC / 4. Februar' 2016
" >>>> Today Atos announced that the French CEA and its industrial partners at the Centre for Computing Research and Technology, CCRT, have invested in a new 1.4 petaflop Bull supercomputer. Three times more powerful than the current computer at CCRT, the new system will be installed in the CEA’s Very Large Computing Centre in Bruyères-le-Châtel, France, mid-2016 to cover expanding industrial needs....
COBALT supercomputer will comprise 2304 Intel Xeon E5 processors (Broadwell) with a total of 32,256 cores at 2.4 Ghz and
18 hybrid nodes with Nvidia Pascal processors, for remote computing and visualization. The partition dedicated to the France Génomique project shall in turn, be equipped with 4032 Intel Xeon E5 Broadwell cores at 2.4 Ghz and 4 very large memory nodes at 3TB/node. An InfiniBand EDR interconnect solution (100 GB/s vs. 40 GB/s currently) forms the core of this new computer and will allow it to increase capacity based on the needs of its partners for the coming 4 to 5 years. <<<< "
Oder siehe hier, einer meiner früheren Post's /// =>
#104 Nvidia Pascal: Fertigung soll in 16 nm FinFET über TSMC laufen
- - - - - - - - - - - -
=> Nvidia/IBM to Build Two Coral 100+ Petaflop Supercomputers in 2017
=> Heterogeneous HPC, Architectural Optimization, and NVLink GPU Interconnect
=> November 17, 2014 / NVIDIA Volta GPU, IBM POWER9 - land Big Contracts for New US Government - Supercomputers in 2017
The Department of Energy has announced
IBM and Nvidia Win $425 Million in Supercomputer Deals, that they will be awarding $425 million in grants to build 100+ petaflop supercomputers using IBM and Nvidia hardware.
- - - - - - - - - - - -
In Anbetracht der riesengrossen Nachfrage & Interesses der ganzen Industrie an Nvidia 'Next-Gen PASCAL' GPU's ( Game Developer / Virtual Reality & Augmented Reality / HPC Supercomputing ) wird man aber sicher ggf. schon einen Monat zuvor auf der diesjährigen GDC - Game Developers Conference im März vom 16th-18th @ San Francisco - schon weitere Einblicke zur neuen Nvidia GPU Architektur gewähren - wenn auch nicht für alle Normalsterblichen, sondern nur wieder hinter verschlossenen Türen ;o)
Nvidia 2 Jahres 'Nick-Nock' - 'Tick-Tock':
Tesla GT200 - 2008
Fermi GF100 - 2010
Kepler GK110 - 2012
Maxwell GM200 - 2014
Next-Gen Pascal GP100 - 2016
Pascal 'refesh' Volta GV100 - 2017