nvidia

Auto Added by WPeMatico

Nvidia’s T4 GPUs are now available in beta on Google Cloud

Published January 16, 2019

Google Cloud today announced that Nvidia’s Turing-based Tesla T4 data center GPUs are now available in beta in its data centers in Brazil, India, Netherlands, Singapore, Tokyo and the United States. Google first announced a private test of these cards in November, but that was a very limited alpha test. All developers can now take these new T4 GPUs for a spin through Google’s Compute Engine service.

The T4, which essentially uses the same processor architecture as Nvidia’s RTX cards for consumers, slots in-between the existing Nvidia V100 and P4 GPUs on the Google Cloud Platform . While the V100 is optimized for machine learning, though, the T4 (as its P4 predecessor) is more of a general-purpose GPU that also turns out to be great for training models and inferencing.

In terms of machine and deep learning performance, the 16GB T4 is significantly slower than the V100, though if you are mostly running inference on the cards, you may actually see a speed boost. Unsurprisingly, using the T4 is also cheaper than the V100, starting at $0.95 per hour compared to $2.48 per hour for the V100, with another discount for using preemptible VMs and Google’s usual sustained use discounts.

Google says that the card’s 16GB memory should easily handle large machine learning models and the ability to run multiple smaller models at the same time. The standard PCI Express 3.0 card also comes with support for Nvidia’s Tensor Cores to accelerate deep learning and Nvidia’s new RTX ray-tracing cores. Performance tops out at 260 TOPS and developers can connect up to four T4 GPUs to a virtual machine.

It’s worth stressing that this is also the first GPU in the Google Cloud lineup that supports Nvidia’s ray-tracing technology. There isn’t a lot of software on the market yet that actually makes use of this technique, which allows you to render more lifelike images in real time, but if you need a virtual workstation with a powerful next-generation graphics card, that’s now an option.

With today’s beta launch of the T4, Google Cloud now offers quite a variety of Nvidia GPUs, including the K80, P4, P100 and V100, all at different price points and with different performance characteristics.

Don’t expect a new Nvidia Shield Tablet anytime soon

Published January 9, 2019

The Shield TV, Nvidia’s Android TV streaming box, is still getting regular updates, but the Shield Tablet, which launched in 2014, was last refreshed in 2015 and officially discontinued last year, wasn’t quite the same success. As Nvidia CEO Jensen Huang said during a small press gathering at CES in Las Vegas today, the company doesn’t have any plans to resurrect it.

“Shield TV is still unquestionably the best Android TV in the world,” he said. “We have updated the software now over 30 times. People are blown away by how much we continue to enhance it.” And more (unspecified) enhancements are coming, he said.

On the mobile side, though, the days of the Shield Tablet are very much over, especially now that the Nintendo Switch, which uses Nvidia’s Tegra chips, has really captured that market.

“We are really committed to [Shield TV], but on mobile devices, we don’t think it’s necessary,” Huang said. “We would only build things not to gain market share. Nvidia is not a ‘take somebody else’s market share company.’ I think that’s really angry. It’s an angry way to run a business. Creating new markets, expanding the horizon, creating things that the world doesn’t have, that’s a loving way to build a business.”

He added that this is the way to inspire employees, too. Just copying competitors and maybe selling a product cheaper, though, does nothing to motivate employees and is not what Nvidia is interested in.

Of course, Huang left the door open to a future tablet if it made sense — though he clearly doesn’t think it does today. He’d only do so, “if the world needs it. But at the moment, I just don’t see it. I think Nintendo did such a great job.”

Bonus: The outspoken Huang also used his time with the assembled journalists to voice his opinion of AMD’s new Radeon VII graphics cards, which were announced earlier today. “Wow. Underwhelming, huh? I was kind of like saying ‘what?’ Because the performance is lousy and there’s nothing new. There’s no raytracing, no artificial intelligence. It’s a 7nm chip with HBM memory that barely keeps up with a 2080 and when we turn on DLSS, we’ll crush it. When we turn on raytracing, we’ll crush it. And it’s not even available yet.”

Nvidia launches Rapids to help bring GPU acceleration to data analytics

Published October 10, 2018

Nvidia, together with partners like IBM, HPE, Oracle, Databricks and others, is launching a new open-source platform for data science and machine learning today. Rapids, as the company is calling it, is all about making it easier for large businesses to use the power of GPUs to quickly analyze massive amounts of data and then use that to build machine learning models.

“Businesses are increasingly data-driven,” Nvidia’s VP of Accelerated Computing Ian Buck told me. “They sense the market and the environment and the behavior and operations of their business through the data they’ve collected. We’ve just come through a decade of big data and the output of that data is using analytics and AI. But most it is still using traditional machine learning to recognize complex patterns, detect changes and make predictions that directly impact their bottom line.”

The idea behind Rapids then is to work with the existing popular open-source libraries and platforms that data scientists use today and accelerate them using GPUs. Rapids integrates with these libraries to provide accelerated analytics, machine learning and — in the future — visualization.

Rapids is based on Python, Buck noted; it has interfaces that are similar to Pandas and Scikit, two very popular machine learning and data analysis libraries, and it’s based on Apache Arrow for in-memory database processing. It can scale from a single GPU to multiple notes and IBM notes that the platform can achieve improvements of up to 50x for some specific use cases when compared to running the same algorithms on CPUs (though that’s not all that surprising, given what we’ve seen from other GPU-accelerated workloads in the past).

Buck noted that Rapids is the result of a multi-year effort to develop a rich enough set of libraries and algorithms, get them running well on GPUs and build the relationships with the open-source projects involved.

“It’s designed to accelerate data science end-to-end,” Buck explained. “From the data prep to machine learning and for those who want to take the next step, deep learning. Through Arrow, Spark users can easily move data into the Rapids platform for acceleration.”

Indeed, Spark is surely going to be one of the major use cases here, so it’s no wonder that Databricks, the company founded by the team behind Spark, is one of the early partners.

“We have multiple ongoing projects to integrate Spark better with native accelerators, including Apache Arrow support and GPU scheduling with Project Hydrogen,” said Spark founder Matei Zaharia in today’s announcement. “We believe that RAPIDS is an exciting new opportunity to scale our customers’ data science and AI workloads.”

Nvidia is also working with Anaconda, BlazingDB, PyData, Quansight and scikit-learn, as well as Wes McKinney, the head of Ursa Labs and the creator of Apache Arrow and Pandas.

Another partner is IBM, which plans to bring Rapids support to many of its services and platforms, including its PowerAI tools for running data science and AI workloads on GPU-accelerated Power9 servers, IBM Watson Studio and Watson Machine Learning and the IBM Cloud with its GPU-enabled machines. “At IBM, we’re very interested in anything that enables higher performance, better business outcomes for data science and machine learning — and we think Nvidia has something very unique here,” Rob Thomas, the GM of IBM Analytics told me.

“The main benefit to the community is that through an entirely free and open-source set of libraries that are directly compatible with the existing algorithms and subroutines that their used to — they now get access to GPU-accelerated versions of them,” Buck said. He also stressed that Rapids isn’t trying to compete with existing machine learning solutions. “Part of the reason why Rapids is open source is so that you can easily incorporate those machine learning subroutines into their software and get the benefits of it.”

Nvidia launches the Tesla T4, its fastest data center inferencing platform yet

Published September 13, 2018

Nvidia today announced its new GPU for machine learning and inferencing in the data center. The new Tesla T4 GPUs (where the ‘T’ stands for Nvidia’s new Turing architecture) are the successors to the current batch of P4 GPUs that virtually every major cloud computing provider now offers. Google, Nvidia said, will be among the first to bring the new T4 GPUs to its Cloud Platform.

Nvidia argues that the T4s are significantly faster than the P4s. For language inferencing, for example, the T4 is 34 times faster than using a CPU and more than 3.5 times faster than the P4. Peak performance for the P4 is 260 TOPS for 4-bit integer operations and 65 TOPS for floating point operations. The T4 sits on a standard low-profile 75 watt PCI-e card.

What’s most important, though, is that Nvidia designed these chips specifically for AI inferencing. “What makes Tesla T4 such an efficient GPU for inferencing is the new Turing tensor core,” said Ian Buck, Nvidia’s VP and GM of its Tesla data center business. “[Nvidia CEO] Jensen [Huang] already talked about the Tensor core and what it can do for gaming and rendering and for AI, but for inferencing — that’s what it’s designed for.” In total, the chip features 320 Turing Tensor cores and 2,560 CUDA cores.

In addition to the new chip, Nvidia is also launching a refresh of its TensorRT software for optimizing deep learning models. This new version also includes the TensorRT inference server, a fully containerized microservice for data center inferencing that plugs seamlessly into an existing Kubernetes infrastructure.

Watch Nvidia unveil the RTX 2080 live right here

Published August 20, 2018

Nvidia is taking advantage of the Gamescom in Germany to hold a press conference about its future graphics processing units. The conference will start at 6 PM in Germany, 12 PM in New York, 9 AM in San Francisco.

Just a week after the company unveiled its new Turing architecture, Nvidia could share more details about the configurations and prices of its upcoming products — the RTX 2080, RTX 2080 Ti, etc.

The name of the conference #BeForeTheGame suggests that Nvidia is going to focus on consumer products and in particular GPUs for gamers. While the GeForce GTX 1080 is still doing fine when it comes to playing demanding games, the company is always working on new generations to push the graphical boundaries of your computer.

According to Next INpact, you can expect two different products this afternoon. The GeForce RTX 2080 is going to feature 2,944 CUDA cores with 8GB of GDDR6. The GeForce RTX 2080 Ti could feature as many as 4,352 CUDA cores with 11GB of GDDR6.

Nvidia already unveiled Quadro RTX models for professional workstations last week. The company is expecting significant performance improvements with this new generation as those GPUs are optimized for ray tracing — the “RT” in RTX stands for ray tracing.

While ray tracing isn’t new, it’s hard to process images using this method with current hardware. The RTX GPUs will have dedicated hardware units for this task in particular.

And maybe it’s going to become easier to buy GPUs now that the cryptocurrency mining craze is slowly fading away.

Microsoft is building low-cost, streaming-only Xbox, says report

Published July 24, 2018

It was revealed at E3 last month that Microsoft was building a cloud gaming system. A report today calls that system Scarlett Cloud and it’s only part of Microsoft’s next-gen Xbox strategy. And it makes a lot of sense, too.

According to Thurrott.com, noted site for all things Microsoft, the next Xbox will come in two flavors. One will be a traditional gaming console where games are processed locally. You know, like how it works on game systems right now. The other system will be a lower-powered system that will stream games from the cloud — most likely, Microsoft’s Azure cloud.

This streaming system will still have some processing power, which is in part to counter latency traditionally associated with streaming games. Apparently part of the game will run locally while the rest is streamed to the system.

The streaming Xbox will likely be available at a much lower cost than the traditional Xbox. And why not. Microsoft has sold Xbox systems with a slim profit margin, relying on sales of games and online services to make up the difference. A streaming service that’s talked about on Thurrott would further take advantage of this model while tapping into Microsoft’s deep understanding of cloud computing.

A few companies have tried streaming full video games. Onlive was one of the first; while successful for a time, it eventually went through a dramatic round of layoffs before a surprise sale for $4.8 million in 2012. Sony offers an extensive library of PS2, PS3 and PS4 games for streaming through its PlayStation Now service. Nvidia got into the streaming game this year and offers a small selection of streaming through GeForce Now. But these are all side projects for the companies.

Sony and Nintendo do not have the global cloud computing platform of Microsoft, and if Microsoft’s streaming service hits, it could change the landscape and force competitors to reevaluate everything.

Reflections on E3 2018

Published June 14, 2018

After taking a year off, I returned to E3 this week. It’s always a fun show, in spite of the fact that the show floor has come to rival Comic-Con in terms of the mass of people the show’s organizers are able to cram into the aisles of the convention center floor.

We’ve been filing stories all week, but here is a very much incomplete collection of my thoughts on this year’s show.

Zombies are still very much a thing

I’d have thought we’d have hit peak zombie years ago, but here we are, zombies everywhere. That includes the LA Convention Center lobby, which was swarming with actors decked out as the undead. There’s something fundamentally disturbing about watching gamers get pictures taken with fake, bloody corpses. Or maybe it’s just the perfect allegory for our time.

Nintendo’s back

A slight adjustment in approach certainly played a role, as the company has embraced mobile gaming. But the key to Nintendo’s return was a refocus on what it does best: offering an innovative experience with familiar IP. Oh, and the GameCube controller Smash Bros. compatibility was a brilliant bit of fan service, even by Nintendo’s standards.

Quantity versus quality?

Microsoft’s event was a sort of video game blitzkrieg. The company showed off 50 titles, a list that included 15 exclusives. Sony, on the other hand, stuck to a handful, but presented them in much greater depth. Ultimately, I have to say I preferred the latter. Real game play footage feels like an extremely finite resource at these events.

Ultra violence in ultra high-def

Certainly not a new trend in gaming, but there’s something about watching someone bite off someone else’s face on the big screen that’s extra upsetting. Sony’s press conference was a strange sort of poetry, with some of the week’s most stunning imagery knee-deep in blood and gore.

Reedus ’n fetus

We saw more footage and somehow we understand the game less?

Checkmate

Indiecade is always a favorite destination at E3. It’s a nice respite from the big three’s packed booths. Interestingly, there were a lot more desktop games than I remember. You know, the real kind with physical pieces and no screens.

Death of a Tomb Raider

I played Shadow of the Tomb Raider on a PC in NVIDIA’s meeting space. It’s good, but I’m not good at it. I killed poor Lara A LOT. I can deal with that sort of thing when my character is in full Master Chief regalia or whatever, but those close-up shots of her face when I drowned her for the fifth time kind of bummed me out. Can video games help foster empathy or are we all just destined to desensitize ourselves because we have tombs to raid, damn it?

I saw the light

NVIDIA also promised me that its ray-tracing tech would be the most impressive demo I saw at E3 that day. I think they were probably right, so take that, Sonic Racing. The tech, which was first demoed at GDC, “brings real-time, cinematic-quality rendering to content creators and game developers.”

VR’s still waiting in the wings

At E3 two years ago, gaming felt like an industry on the cusp of a VR breakthrough. In 2018, however, it doesn’t feel any closer. There were a handful of compelling new VR experiences at the event, but it felt like many of the peripheral and other experiences were sitting on the fringes of the event — both literally and metaphorically — waiting for a crack at the big show.

Remote Control

Sony’s Control trailer was the highest ratio of excitement to actual information I experienced. Maybe it’s Inception the video game or the second coming of Quantum Break. I dunno, looks fun.

AR’s a thing, but not, like, an E3 thing

We saw a few interesting examples of this, including the weirdly wonderful TendAR, which requires you to make a bunch of faces so a fake fish doesn’t die. It’s kind of like version of Seaman that feeds on your own psychic energy. At the end of the day, though, E3 isn’t a mobile show.

Cross-platform

Having said that, there are some interesting examples of cross-platform potential popping up here and there. The $50 Poké Ball Plus for the Switch is a good example I’m surprised hasn’t been talked about more. Along with controlling the new Switch titles, it can be used to capture Pokémon via Pokémon GO. There’s some good brand synergy right there. And then, of course, there’s Fortnite, which is also on the Switch. The game’s battle royale mode is a great example of how cross-platform play can lead to massive success. Though by all accounts, Sony doesn’t really want to play ball.

V-Bucks

Oh, Epic Games has more money than God now.

Moebius strip

Video games are art. You knew that already, blah, blah, blah. But Sable looks like a freaking Moebius comic come to life. I worry that it will be about as playable as Dragon’s Lair, but even that trailer is a remarkable thing.

Nvidia launches colossal HGX-2 cloud server to power HPC and AI

Published May 30, 2018

Nvidia launched a monster box yesterday called the HGX-2, and it’s the stuff that geek dreams are made of. It’s a cloud server that is purported to be so powerful it combines high-performance computing with artificial intelligence requirements in one exceptionally compelling package.

You know you want to know the specs, so let’s get to it: It starts with 16x NVIDIA Tesla V100 GPUs. That’s good for 2 petaFLOPS for AI with low precision, 250 teraFLOPS for medium precision and 125 teraFLOPS for those times when you need the highest precision. It comes standard with a 1/2 a terabyte of memory and 12 Nvidia NVSwitches, which enable GPU to GPU communications at 300 GB per second. They have doubled the capacity from the HGX-1 released last year.

Chart: Nvidia

Paresh Kharya, group product marketing manager for Nvidia’s Tesla data center products, says this communication speed enables them to treat the GPUs essentially as a one giant, single GPU. “And what that allows [developers] to do is not just access that massive compute power, but also access that half a terabyte of GPU memory as a single memory block in their programs,” he explained.

Graphic: Nvidia

Unfortunately you won’t be able to buy one of these boxes. In fact, Nvidia is distributing them strictly to resellers, who will likely package these babies up and sell them to hyperscale data centers and cloud providers. The beauty of this approach for cloud resellers is that when they buy it, they have the entire range of precision in a single box, Kharya said.

“The benefit of the unified platform is as companies and cloud providers are building out their infrastructure, they can standardize on a single unified architecture that supports the entire range of high-performance workloads. So whether it’s AI, or whether it’s high-performance simulations, the entire range of workloads is now possible in just a single platform,”Kharya explained.

He points out this is particularly important in large-scale data centers. “In hyperscale companies or cloud providers, the main benefit that they’re providing is the economies of scale. If they can standardize on the fewest possible architectures, they can really maximize the operational efficiency. And what HGX allows them to do is to standardize on that single unified platform,” he added.

As for developers, they can write programs that take advantage of the underlying technologies and program in the exact level of precision they require from a single box.

The HGX-2 powered servers will be available later this year from partner resellers, including Lenovo, QCT, Supermicro and Wiwynn.

Gaming monitors, headsets and peripherals for a winning desktop setup

Published May 23, 2018

Makula Dunbar
Contributor

Makula Dunbar is a writer with Wirecutter.

Pure Storage teams with Nvidia on GPU-fueled Flash storage solution for AI

Published March 27, 2018

As companies gather increasing amounts of data, they face a choice over bottlenecks. They can have it in the storage component or the backend compute system. Some companies have attacked the problem by using GPUs to streamline the back end problem or Flash storage to speed up the storage problem. Pure Storage wants to give customers the best of both worlds.

Today it announced, Airi, a complete data storage solution for AI workloads in a box.

Under the hood Airi starts with a Pure Storage FlashBlade, a storage solution that Pure created specifically with AI and machine learning kind of processing in mind. NVidia contributes the pure power with four NVIDIA DGX-1 supercomputers, delivering four petaFLOPS of performance with NVIDIA ® Tesla ® V100 GPUs. Arista provides the networking hardware to make it all work together with Arista 100GbE switches. The software glue layer comes from the NVIDIA GPU Cloud deep learning stack and Pure Storage AIRI Scaling Toolkit.

Photo: Pure Storage

One interesting aspect of this deal is that the FlashBlade product operates as a separate product inside of the Pure Storage organization. They have put together a team of engineers with AI and data pipeline understanding with the focus inside the company on finding ways to move beyond the traditional storage market and find out where the market is going.

This approach certainly does that, but the question is do companies want to chase the on-prem hardware approach or take this kind of data to the cloud. Pure would argue that the data gravity of AI workloads would make this difficult to achieve with a cloud solution, but we are seeing increasingly large amounts of data moving to the cloud with the cloud vendors providing tools for data scientists to process that data.

If companies choose to go the hardware route over the cloud, each vendor in this equation — whether Nvidia, Pure Storage or Arista — should benefit from a multi-vendor sale. The idea ultimately is to provide customers with a one-stop solution they can install quickly inside a data center if that’s the approach they want to take.