NVIDIA founder and CEO Jensen Huang sent the following letter to our enlarged employee base when the Mellanox acquisition closed earlier this week: Hi everyone, I hope all of you and your families are well and safe during these extraordinary times. On March 11 last year, we announced plans to buy Mellanox, the world leader in Read article >
The post Letter From Jensen: Building the New NVIDIA Together appeared first on The Official NVIDIA Blog.
NVIDIA founder and CEO Jensen Huang sent the following letter to our enlarged employee base when the Mellanox acquisition closed earlier this week:
I hope all of you and your families are well and safe during these extraordinary times.
On March 11 last year, we announced plans to buy Mellanox, the world leader in high-performance networking. Last week, China gave us the green light. In addition to the U.S., E.U., and Mexico, all jurisdictions have approved our combination. I am thrilled to announce that today, April 27, 2020, NVIDIA and Mellanox are officially one company!
Mellanox is now NVIDIA’s networking brand, our networking business unit, and Israel is one of NVIDIA’s most important technology centers, with 2,500 employees. The new NVIDIA has a much larger footprint in data center computing, end-to-end and full-stack expertise in data center architecture, and tremendous scale to accelerate innovations. We are uniting two powerful forces and doubling down on data center.
Our journey started in my hotel room at SC18 on November 18, 2018. Eyal and I had a 7 am breakfast meeting at the Dallas Ritz-Carlton. He told me there were multiple parties interested in buying Mellanox and asked if NVIDIA would like to join the bid. I told him that I loved Mellanox – the company, the technology, the products, and strategic position. We’ve partnered to build many supercomputers and AI systems. “Mellanox is an amazing company,” I said to Eyal. He went on for the next hour, telling me that “Mellanox is an amazing company.” I agreed. He’s incredibly proud of Mellanox, and rightfully so. At the end of the conversation, I told him that NVIDIA would be the final and highest bidder. The bidding was intense. NVIDIA was the highest bidder.
NVIDIA and Mellanox are a perfect combination. The first reason is that we both care deeply about high-performance computing. Our companies are both wired for performance and technology leadership. NVIDIA accelerates 50% of the new supercomputers on the latest Top500 list. Mellanox is the networking in 6 of the top 10 supercomputers. When NVIDIA accelerates compute by 10-50x, moving data becomes the bottleneck. Amdahl’s law. Mellanox and NVIDIA amplify each other’s technologies.
The big picture reason is that the future of the data center is being reshaped and playing into our strengths. Yesterday’s cloud data center architecture is called Hyper-Converged Infrastructure, where CPU-servers with integrated storage “scale out” into large data centers. Cloud computing was born and revolutionized the industry. AI is now driving an architecture change from Hyper-Converged Infrastructures to Accelerated-Disaggregated Infrastructure. NVIDIA and Mellanox are at the epicenter of this shift.
AI, the most powerful technology force of our time, is a fundamentally new computing model that is reshaping data center architectures. AI is insanely computation- and data-intensive. Because CPUs designed for general-purpose computing cannot efficiently handle the workload, NVIDIA is used to accelerate AI algorithm processing, and Mellanox is used to accelerate data processing and networking. The data center is now accelerated.
Simultaneously, the rapid adoption of data center-scale workload orchestration platforms like Kubernetes is changing software development and data center operations in a significant way. Applications and services are decomposed into smaller executable containers and deployed onto available servers across the data center. Services can dynamically compose the infrastructure with the optimal type and amount of computing elements to process each workload. Large monolithic applications no longer bog down some computers while leaving others sitting idle. Data center utilization, throughput, ease and cost of operations are all improved.
The data center is now a composable computer. The data center computing elements are disaggregated — i.e., CPU servers, GPU servers, and storage servers are separated and scaled independently. Mellanox’s networking fabric that connects the computing elements becomes all-important to the data center.
Our growing businesses reflect the need for fast computing and networking. When we announced the acquisition a year ago, Mellanox did $305M in revenue that quarter. Mellanox had 4 consecutive record quarters since and just announced $429M in quarterly revenue – increasing 41%. During that same period, NVIDIA’s data center business grew from $680M to approaching $1B last quarter, also increased over 40%. This is just the beginning.
Breakthroughs in natural language understanding, conversational AI, and deep recommendation systems are pushing AI complexity to new levels. The models are complex. Data intensity is growing exponentially. And the relentless push to smarter predictions and faster response is driving computation demands.
AI computing, networking, and data processing are more important than ever in this new world. Together we possess deep expertise and exceptional capabilities to invent this future. NVIDIA and Mellanox are a dream team combination.
The fastest computing and the fastest networking go hand in hand. That’s why NVIDIA selected Mellanox for our DGX AI systems. Mellanox switches and networking fabric connect NVIDIA’s own research AI supercomputers. Mellanox data processing chips do networking, storage, and security for our EGX edge AI and 5G platform and our AGX self-driving car computer.
It is a tremendous joy to work with Mellanox’s world-class people. Getting to know you more deeply this past year, we love the company, the technology, the products, and strategic position even more. Eyal, the management team, and all the people of Mellanox have built a truly extraordinary company.
We are thrilled to become one family and invent the future together.
Bruchim abahim le mishpachat NVIDIA!
The post Letter From Jensen: Building the New NVIDIA Together appeared first on The Official NVIDIA Blog.
Seagate Technology ships tens of millions of hard disk drives every quarter. Ensuring the quality of each one is a top priority, but not easy. The disk drive manufacturing process is incredibly complex. For example, it takes 1,400 steps just to manufacture the drive head. Even the smallest errors can lead to product flaws. “Mistakes Read article >
The post Seagate Provides the RX to Improve Hard Drive Manufacturing appeared first on The Official NVIDIA Blog.
Seagate Technology ships tens of millions of hard disk drives every quarter. Ensuring the quality of each one is a top priority, but not easy.
The disk drive manufacturing process is incredibly complex. For example, it takes 1,400 steps just to manufacture the drive head. Even the smallest errors can lead to product flaws.
“Mistakes are costly,” said Raghavan Srinivasan, senior director of global marketing at Seagate, which makes more disk drives than any other manufacturer. “If there are any anomalies, we’re going to end up with poor yield later on, and we wouldn’t know that until the long process is completed.”
To combat that, Seagate is using GPU-powered AI and machine learning to perform anomaly detection of recording head slider images so it can spot potential problems sooner.
A Solution Born from Partnership
At first, Seagate searched for a product that could help with image-based anomaly detection in a high-yield manufacturing environment, but there was nothing commercially available. The company started talking with NVIDIA and Hewlett Packard Enterprise about the business problem.
The first step the companies took was to get an idea of the amount of data in play. Bringing this idea to life for a single factory would require the ability to process 17 million images a day, and to perform inference at 20 positions per second.
“It’s like a [Boeing] 747 flying inches from the ground at 100 times the speed of sound,” said Srinivasan.
Over the course of a year, the companies built a solution that could capture and analyze images of drive heads that are just molecules from the surface of the spinning disk.
Seagate deployed this solution in a machine vision defect inspection system used in hard disk read-and-write head manufacturing in one of Seagate’s factory sites and is planning to scale the solution to other sites.
Because the anomaly patterns in question were constantly changing, rules-based AI was not an option. A deep learning approach was needed.
What’s more, Seagate knew the cost of bandwidth to move so much data to and from the factory would be prohibitive, so it developed a parallel reference architecture called Seagate Edge RX. This architecture establishes each factory as a computing “edge,” enabling the inference process to occur on the factory floor, using lighter-weight computing resources.
Ongoing training of the deep learning model occurs offline and is performed by data scientists, who then deliver updated models.
As this technology becomes incorporated into all of Seagate’s manufacturing processes,
Seagate expects to see up to a 20 percent reduction in cleanroom investments, a 10 percent reduction in manufacturing throughput time, and up to a 300 percent return on investment from improved efficiency and better quality.
Srinivasan said there’s every indication that the project will ultimately yield significant improvements in quality, providing plenty of juice for expanding the effort to all of Seagate’s manufacturing facilities.
It’s not lost on Seagate that none of this would’ve been possible without technologies from NVIDIA and HPE. These include rugged HPE Edgeline systems with NVIDIA T4 GPUs for data acquisition and real-time AI inferencing at the factory floor, and HPE Apollo systems with NVIDIA V100 Tensor Core GPUs for AI training. The NVIDIA EGX platform runs on these systems and allows enterprises to securely deploy and manage AI workloads at the edge or in the data center.
“NVIDIA has become synonymous with AI and deep learning, and being able to provide the process behind it,” said Srinivasan.
Seagate sees a lot of potential to take this work in a variety of directions. For now, the company is focused on rolling out the smart manufacturing platform to all of its manufacturing sites. It also plans to explore how the solution can impact other manufacturing processes. Down the line, Seagate will explore predictive maintenance of process tools based on IoT sensor-based logging.
The good news for other manufacturers is that Seagate doesn’t want to keep all of this innovation to itself. The company has published the Seagate Edge RX parallel reference architecture to give other manufacturers a jump-start in building similar image anomaly detection solutions in their own manufacturing environments.
“We want to move the conversation forward,” said Srinivasan. “We see there are a lot more opportunities to improve overall manufacturing if you put data to work.”
Learn more about Seagate’s work with NVIDIA in an on-demand talk on AI on EGX: Innovations and Lessons Learned from a Production Deployment Across Global Manufacturing Sites available now with a free registration to GTC Digital.
The post Seagate Provides the RX to Improve Hard Drive Manufacturing appeared first on The Official NVIDIA Blog.
» Download all images (ZIP, 14 MB)
What’s New: Intel introduced today the 10th Gen Intel® Core™ S-series desktop processors, including Intel’s flagship Core i9-10900K processor, the world’s fastest gaming processor1. With speeds reaching up to a maximum of 5.3 GHz with Intel® Thermal Velocity Boost2 out of the box, 10th Gen Intel Core desktop processors deliver real-world performance for a new level of experience in gaming.
“Intel is committed to enabling the future of powerful desktop gaming by continuously pushing the performance boundary to deliver an amazing PC gaming experience. The 10th Gen Intel Core S-series for desktops, and the Intel Core i9-10900K processor, the world’s fastest gaming processor reinforces our commitment to the gaming and enthusiast communities.”
– Brandt Guttridge, Intel senior director of the Desktop Products Group
Why It’s Great for Gaming and Overclocking: At the top of the stack is the unlocked 10th Gen Intel Core i9-10900K, the world’s fastest gaming processor1, featuring up to 10 cores, 20 threads and DDR4-2933 memory speeds. The i9-10900K processor powers the ultimate gaming experience, allowing more tuning control, faster multitasking and smoother gameplay. The new Intel® Turbo Boost Max Technology 3.0 provides automatic performance boosts on lightly threaded applications, while per-core hyperthreading control allows experienced overclockers to decide which threads to turn on or off on a per-core basis. Improvements in this generation include:
- Up to 187 frames per second3 for in-game performance while streaming and recording, and up to 63 percent more frames per second4 in gaming compared with a 3-year-old PC.
- Up to 12 percent faster5 video editing compared with the previous generation, and up to 15 percent faster6 video editing compared with a 3-year-old PC.
- Up 18 percent faster7 4K video editing compared with the previous generation, and up to 35 percent faster8 4K video editing compared with a 3-year-old PC.
- Up to 31 percent better overall system performance compared with a 3-year-old PC9, 10.
About Key Features and Capabilities: The 10th Gen Intel Core S-series processors deliver smooth gameplay through best-in-class connectivity11, immersive entertainment and enhanced streaming.
- Intel Thermal Velocity Boost: Gamers and creators get an opportunistic and automatic boost across single-core and multicore workloads, with up to 5.3 GHz.
- Intel Hyperthreading Technology: Now across 10th Gen Intel Core i9 to i3 processors.
- Enhanced Core & Memory Overclocking3: Gain ultimate control when overclocking your processor and key system components with features enabled by new unlocked and overclockable 10th Gen Intel Core processors.
- Intel® Ethernet Connection I225: Now available on the 10th Gen platform, 2.5G Intel Ethernet Connector I225 delivers greater than two times the network speeds of 1GB Ethernet on existing cabling.
- Intel® Wi-Fi 6 AX201: Now integrated into 10th Gen Intel Core desktop processors, Intel Wi-Fi 6 (Gig+) delivers responsive gameplay, nearly three times faster12 downloads and more reliable connections. Delivers best-in-class connectivity11 with the freedom and flexibility to game or create anywhere throughout the home or office.
When You Can Get It: 10th Gen Intel Core S-series processors are expected to be available globally through normal retail channels and in desktops sold worldwide by OEMs and channel system integrators, starting in May.
More Context: 10th Gen Intel Core S-series (Product Brief) | 10th Gen Intel Core Processors – Amazing Gaming Starts With Intel (Press Kit) | More Powerful Edge Processing with 10th Generation Intel Core Processors for IoT and Edge | A New Power Standard has Desktops Sipping Energy when Idling | Intel® Z490 Chipset (Product Brief)
The Small Print: All information provided here is subject to change without notice. Contact your Intel representative to obtain the latest Intel product specifications, roadmaps, and related information.
Performance results are based on testing as of April 17, 2020 in configurations and may not reflect all publicly available updates. See backup for configuration details. No product or component can be absolutely secure.
Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors. Performance tests, such as SYSmark and MobileMark, are measured using specific computer systems, components, software, operations and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products. For more complete information about performance and benchmark results, visit http://www.intel.com/benchmarks
Intel technologies may require enabled hardware, software or service activation.
Your costs and results may vary.
Statements in this document that refer to Intel’s plans and expectations for the quarter, the year, and the future, are forward-looking statements that involve a number of risks and uncertainties. A detailed discussion of the factors that could affect Intel’s results and plans is included in Intel’s SEC filings, including the annual report on Form 10-K.
Altering clock frequency or voltage may void any product warranties and reduce stability, security, performance, and life of the processor and other components. Check with system and component manufacturers for details.
© Intel Corporation. Intel, the Intel logo, and other Intel marks are trademarks of Intel Corporation or its subsidiaries. Other names and brands may be claimed as the property of others.
1As measured by in-game benchmark mode performance (score or frames per second) where available, or frames per second where benchmark mode is unavailable. PC Gaming Processors Compared: 10th Gen Intel® Core™ i9-10900K, Intel® Core™ i9-9900KS, AMD Ryzen™ 9 3950X. Prices of compared products may differ. Configurations: Graphics: Nvidia GeForce RTX 2080 Ti, Memory: 4x8GB DDR4 (2666, 2933 or 3200 per highest speed of the corresponding processor), Storage: Intel Optane SSD 905P, OS Windows 10 Pro 1909 v720 19H2(RS6). Results: 10th Gen Intel® Core™ i9-10900K scored better on the majority of the 25+ game titles tested. Intel will be marketing the 10th Gen Intel® Core™ i9-10900K with the tag line “Elite Real World Performance” in certain jurisdictions, including PRC and Vietnam. Intel will be marketing the 10th Gen Intel® Core™ i9-10900K with the tag line “Intel’s Fastest Gaming Processor” in certain jurisdictions, including Argentina, Belarus, Belize, Chile, Egypt, El Salvador, Guatemala, Honduras, Italy, Japan, Panama, Peru, Saudi Arabia, Turkey, Russia, and Ukraine. If you are media, an influencer or a marketer from these countries, or otherwise communicating directly to residents in these countries (e.g., on local-language social media), please only refer to the tag line Intel will be using in that country in lieu of the claim on this slide/document.
2Includes the effect of Intel® Thermal Velocity Boost (Intel® TVB), a feature that opportunistically and automatically increases clock frequency above single-core and multi-core Intel® Turbo Boost Technology frequencies based on how much the processor is operating below its maximum temperature and whether turbo power budget is available. The frequency gain and duration is dependent on the workload, capabilities of the processor and the processor cooling solution.
3As measured by Megatasking workload on PUBG on 10th Gen Intel® Core™ i9-10900K.
4As measured by Players Unknown Battleground (PUBG) on 10th Gen Intel® Core™ i9-10900K vs. 7th Gen Intel® Core™ i7-7700K.
5As measured by Adobe Lightroom Classic workload RUG 1010 on Intel® Core™ i9-10900K vs. Intel® Core™ i9- 9900K.
6As measured by Adobe Lightroom Classic workload RUG 1010 on Intel® Core™ i9-10900K vs. Intel® Core™ i7-7700K.
7As measured by Adobe Premier Pro CC workload RUG 1209 on Intel® Core™ i9-10900K vs. Intel® Core™ i9- 9900K.
8As measured by Adobe Premier Pro CC workload RUG 1209 on Intel® Core™ i9-10900K vs. Intel® Core™ i7-7700K.
9As measured by SYSMark*2018 on on Intel® Core™ i9-10900K vs. Intel® Core™ i7-7700K.
10Intel technologies may require enabled hardware, software or service activation.
11‘Best in Class Wi-Fi 6’: Intel® Wi-Fi 6 (Gig+) products support optional 160 MHz channels, enabling the fastest possible theoretical maximum speeds (2402 Mbps) for typical 2×2 802.11ax PC Wi-Fi products. Premium Intel® Wi-Fi 6 (Gig+) products enable 2-4X faster maximum theoretical speeds compared standard 2×2 (1201 Mbps) or 1×1 (600 Mbps) 802.11ax PC Wi-Fi products, which only support the mandatory requirement of 80 MHz channels. Gigabit Wi-Fi Requirements: To achieve speed of over 1 Gbps requires Gig internet service, router/gateway with either Wi-Fi 6 or 11ac with 160 MHz channel support, and PC with Intel® Wireless 9260/9560 or Intel® Wi-Fi 6 (Gig+) AX200/AX201.
12WiFi 6 ‘Nearly 3X Faster’: 802.11ax 2×2 160MHz enables 2402Mbps maximum theoretical data rates, ~3X (2.8X) faster than standard 802.11ac 2×2 80MHz (867Mbps) as documented in IEEE 802.11 wireless standard specifications, and require the use of similarly configured 802.11ax wireless network routers.
» Download all images (ZIP, 7 MB)
Our expert: Stephen Eastman, platform power specialist in Intel’s Client Computing Group, and his team work on reducing the idle power that desktop PC’s use. Desktop computers often spend more time idle than in any other use. Any way to reduce power used during these long stretches of minimal activity is vital to cutting a PC’s overall power consumption.
Less is more: Using less power is always good, but it will be even more important when governments around the world introduce new energy regulations for desktop computers. This includes EPA’s ENERGY STAR for Computers v8 in 2020, the California Energy Commission Title 20, Tier 2 requirement in 2021 (with other U.S. states following), and Japan’s Top Runner program in 2022. All these regulations are requiring PC makers to adjust their desktop power targets.
A single rail power supply design is one answer to help OEMs reduce desktop PC idle power and meet the new government regulations. For years there have been custom single rail power supply designs. But until Intel created the ATX12VO design and publicly shared it, there was no industry standard to help PC-makers reduce the energy desktop systems use when idling.
But it’s more complicated than just a new power supply. The new ATX12VO standard requires a redesign to the motherboard. The two parts work in tandem under the new standard.
“This is a part of Intel’s greater PC story to enable the most energy efficient desktop computers,” says Eastman. Intel is a processor company, he says, but it is also a company that helps OEMs. And that’s why Intel spent the last two years creating and sharing this standard.
The industry’s first ATX12VO motherboard: On April 30, PC manufacturer ASRock launched the industry’s first ATX12VO motherboard around the Intel 10th Gen Intel Core S-series platform. It reduces idle power by 27% compared to a similar featured motherboard and traditional ATX multi-rail power supply design.
“We’re extremely delighted to co-develop this revolutionary motherboard — Z490 Phantom Gaming 4SR — with Intel. This new design is able to improve power efficiency of the PC and meet new energy regulations. We believe the new ATX12VO will be the solution for the next generation of personal computer,” says Chris Lee, general manager of ASRock Motherboard BU.
Intel is also working with power supply unit builders such as FSP, High Power, Channel Well Technology and Corsair to build ATX12VO power supply units to go with the motherboards.
Knowledge is power: Intel is releasing the spec to provide the industry with a consistent design. It’s especially important for medium to smaller OEMs that don’t have the resources to create their own power supply spec. It will allow them to continue selling devices once new energy regulations go into effect. For a consumer who wants to build their own PC, they’ll have to have both the new power supply and the new motherboard that work together.
Currently, power supplies change the AC current at the plug to the DC current your computer needs. That conversion can cause the greatest loss of power when a computer is idling and at its most inefficient. Most power supply units have 12-, 3.3- and 5-volt rails, or circuitry that creates the voltage. The ATX12VO standard takes out the 3.3- and 5-volt rails and moves the creation of these voltages to the motherboard where they can be more energy-efficient.
“This makes no difference to performance. It makes no difference to the end user except for using less idle power. It just makes a difference to the system builder,” Eastman says.
The post A New Power Standard has Desktops Sipping Energy when Idling appeared first on Intel Newsroom.
While many of us have been busy social distancing, OpenBSD development work has continued. Noteworthy things not previously reported here include:
- The OpenBSD version has moved to 6.7-beta
- Some 11 syscalls have been unlocked since the 6.6 release.
- FFS2 has been made the default filesystem for new installs on most platforms.
- The rpki-client web site has been launched.
- Supported hardware on the arm64 platform has widened further, including support for Pine64 Pinebook Pro and Rasperry Pi 4.
- The default compiler on the macppc platform has been switched to clang(1).
- Ports work has entered slowdown in the move towards release.
Dating apps may get all the press, but NerdWallet has been refining the art of financial matchmaking for more than a decade. The company provides its members with sound financial advice generated by machine learning algorithms. But as computing has advanced, so has NerdWallet’s ambition. Now the company is using AI to better match casual Read article >