NVIDIA Planning To Ditch Maxwell GPUs For HPC Purposes Due To Lack of DP Hardware - Will Update Tesla Line With Pascal in 2022, Volta Arriving in 2022

Some major reports on NVIDIA'due south GPU evolution process has arrived from 4Gamer (via Kitguru) which shows what NVIDIA plans to do with their electric current generation Maxwell plus the next generation Pascal and Volta GPU cores. The details come up from Kenichi Hayashi (NVIDIA platform business headquarters Director) who revealed the development roadmap for NVIDIA during the upcoming years at a conference held in Tokyo, Japan.

NVIDIA Maxwell GPUs Ditched For Loftier Performance Computing In Tesla Parts

Do note start that this news comes straight from NVIDIA's official representative so its highly likely that this news is true and NVIDIA would possibly provide an update on this matter during their CES conference when their event official kicks off early on next month. Now according to 4Gamer'due south interview with Mr.Kenichi, it's said that NVIDIA won't opt to employ Maxwell for their Tesla GPU accelerators which are aimed towards the HPC market. While the addition of Maxwell GPUs in the Quadro line of cards would be seen over the upcoming months, those aiming for super calculating won't get an updated Maxwell core and might have to stick with current Kepler based Tesla offerings.

Just over a calendar month back, we saw NVIDIA update their Tesla lineup with the Kepler GK210 based Tesla K80 Accelerator which featured 24 GB GDDR5 memory and effectually eight.v TFlops of single precision performance with its 4992 CUDA Cores. The bill of fare is a dual core Tesla offering and remains to exist seen if NVIDIA will now launch a K80X to replace information technology in the future during the 2022 timeline in response to AMD's own FirePro dual flake offering which is nonetheless to be launched.

The reason behind not deploying Maxwell in Tesla Accelerators is said to be the lack of FP64 Floating Point Units and boosted double precision hardware. And the reason behind non including DP FPUs in Maxwell might accept to exercise with the extremely efficient design that NVIDIA was aiming for. This however ways that NVIDIA'southward upcoming Maxwell core, the GM200 which is the flagship core of the series might only remain the GeForce merely offering unlike Kepler which was aimed for the HPC marketplace first with the Titan supercomputer and launched a year later after the arrival of the initial Kepler cores as the GeForce GTX Titan. While this may sound negative for the HPC market place since they'll take to wait for some other year and a half to get the adjacent GPU based Tesla office, it's good news for the consumer market. it should also be noted that the advantage of owning a consumer level bill of fare with full DP support such as the Titan and Titan Blackness volition be lost.

Since the DP FP64 FPU hardware blocks will be removed from the top-tier cards that are rumored to get in adjacent year, they volition include several more FP32 FPUs to make utilize of the remaining cadre infinite and that means improve functioning for the GeForce parts since games have little to do with Double precision performance. So with no Tesla parts based on Maxwell, HPC and the supercomputing market place will have to practice with either the current GPU parts or wait till adjacent twelvemonth when Pascal arrives. There'south also a slight bit of an update on Pascal and Volta besides.

NVIDIA Ramps Upwards Pascal and Volta Production - Next Generation Volta Arrives in 2022 To HPC Kickoff

We have already heard nigh Volta coming in 2022 during the announcement of the two new supercomputers, the Summit from Oak Ridge National Laboratory and Sierra from Lawrence Livermore National Laboratory. This news just makes it much more official since its coming from Mr.Kenichi. While Pascal would replace Maxwell as the next generation node based on the 16FF+ process node, Volta will be launched in merely a years lifespan of Pascal to update the next supercomputers with an enormous update in compute performance. While it is piece of cake to believe that Pascal will stick with the 2 years lifespan of every new GPU at the consumer level, the HPC side will see an update ane year earlier just after Pascal launches. In 2022, NVIDIA will ship both Laboratories with their new Tesla accelerators based on the Volta GPU compages that would be based presumably on the 10nm FinFET node.

Rated at a peak performance of 150-300 PFLOPS, the Superlative supercomputer will exist based on more 3400 compute nodes with each node consistin of several next generation IBM POWER9 CPUs and NVIDIA Volta compages based Tesla GPUs. Each node will deliver around forty TFLOPs of compute performance and is said to be plenty to outperform an entire rack of summit of the line Haswell based x86 CPU servers. The near impressive characteristic nearly Pinnacle will be that while information technology will swallow x% more power than the Titan supercomputer, it will as well deliver an impeccable five times more system to application power than the previous fastest supercomputer. While Titan was rated at 25-30 PETAFLOPs, Sierra will be >100 PFlops while Summit volition be 150-300 PFlops at whatsoever given circumstance.

With Top come up a few other enabling technologies such as NVLINK which is expected to debut around 2022. The Pascal GPU would be the kickoff to introduce NVLINK which is the next generation Unified Virtual Memory link with Gen 2.0 Cache coherency features and 5 – 12 times the bandwidth of a regular PCIe connection. This volition solve many of the bandwidth issues that high performance GPUs currently confront.

Kickoff technology we'll denote today is an important invention called NVLink. It's a chip-to-chip communication channel. The programming model is PCI Limited but enables unified retention and moves 5-12 times faster than PCIe. "This is a large jump in solving this bottleneck," Jen-Hsun says. NVIDIA

While Pascal volition aim to add together 3D Stacked Memory designs with on-board and on-flake level integration, Volta will further refine this technology and support the new NVLINK tech which is the next generation Unified Virtual Retentivity link with Gen 2.0 Cache coherency features and 5 – 12 times the bandwidth of a regular PCIe connection. This will solve many of the bandwidth issues that high performance GPUs currently face. These supercomputers will alone amount to a major acquirement generation for NVIDIA however the GeForce department still amounts to a bulk of NVIDIA sales, what's going to happen to GeForce parts during the 2022-2018 timeline.

Well, this would be mostly speculation just I believe that 2022 will encounter the arrival of several new GeForce and Quadro parts based on the Maxwell chip architecture. Currently we know only most the GeForce GTX 960 which is allegedly launching on 22nd January and aiming for the sweet price point of $200 US. Aside from that is all just dust but we tin make a blueprint from here. The GM200 volition arrive afterwards in 2022 prior to a few GM204 Quadro parts. In 2022, we volition see NVIDIA update the Telsa market as quickly every bit possible while launching a few GP*04 Pascal parts. The flagship parts would arrive in 2022 near definitely for consumers however they will stick the long run while NVIDIA makes a quick shift towards Volta accelerators the same year. It is unknown when Volta makes fashion to consumers, the appointment could probably be set in 2022 all the same nosotros can't yet confirm since at that place's no official information on that. Here's hoping that NVIDIA provides an update on their GPU road map at CES 2022 in the upcoming week.

Source: https://wccftech.com/nvidia-planning-ditch-maxwell-gpus-hpc-purposes-due-lack-dp-hardware-update-tesla-line-pascal-2016-volta-arriving-2017/

Posted by: shirleyrequed.blogspot.com

Share this post