<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Intel Tech on Medium]]></title>
        <description><![CDATA[Stories by Intel Tech on Medium]]></description>
        <link>https://medium.com/@intel.author?source=rss-a5e564ccb3ea------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*9DNNoTDJ6II3sbztzCWGsg.jpeg</url>
            <title>Stories by Intel Tech on Medium</title>
            <link>https://medium.com/@intel.author?source=rss-a5e564ccb3ea------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 06:17:58 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@intel.author/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[What are Some of the Technologies Powering Smart Cities?]]></title>
            <link>https://medium.com/intel-tech/what-are-some-of-the-technologies-powering-smart-cities-today-e664e084d6f0?source=rss-a5e564ccb3ea------2</link>
            <guid isPermaLink="false">https://medium.com/p/e664e084d6f0</guid>
            <category><![CDATA[mobileye]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[intel]]></category>
            <category><![CDATA[million-girls-moonshot]]></category>
            <category><![CDATA[mindteck]]></category>
            <dc:creator><![CDATA[Intel Tech]]></dc:creator>
            <pubDate>Tue, 14 Dec 2021 13:39:11 GMT</pubDate>
            <atom:updated>2021-12-14T13:39:36.231Z</atom:updated>
            <content:encoded><![CDATA[<h4>Million Girls Moonshot</h4><h4>Authors: David Allison, Writer, Teacher, Coach, and Tech Enthusiast &amp; Justin Tucker, JANA, Inc. Technical Blogger</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/943/1*P6RkKFaUEq7iIVHAled1GA.png" /></figure><p>According to <a href="https://www.un.org/development/desa/en/news/population/2018-revision-of-world-urbanization-prospects.html">UN projections</a>, 55% of the world population resides in urban areas today, with that percentage expected to increase to 68% by 2050. This means cities will become denser while relying on outdated infrastructure that won’t be able to handle these stark growing pains.</p><p>To ensure urban centers can support their growing communities, policymakers and businesses need to come together to devise a solution. At Intel, we’re developing new technologies to make cities smarter to not only support the influx of new urbanites but to sustain it as well.</p><p>Some of the key technologies powering smart cities are the new generation wireless (5G + Wi-Fi) networks, Internet of Things (IoT), and Artificial Intelligence (AI).</p><p>Together, common problems like traffic congestion, public safety, environmental sustainability, and healthcare can be remedied to improve the lives of city-dwellers and avoid economic pitfalls. Take a look at how Intel is improving urban lives with smart city technologies.</p><h4>Healthcare</h4><p>If there’s one thing 2020 has taught us, it’s that healthcare is essential. As the COVID-19 pandemic continues to sweep across the globe, large cities are finding themselves in troubled waters trying to keep their communities safe. Intel’s <a href="https://newsroom.intel.com/news/intel-commits-technology-response-combat-coronavirus/">Pandemic Response Technology Initiative</a> was implemented to help bring access to online learning for students, speed scientific research, and bring access to technology at the point of patient care.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*4bvBJh-PL5sqR3T7-UCnag.png" /></figure><p>In diverse <a href="https://newsroom.intel.com/news/houston-intel-tech-healthcare-education-connectivity/#gs.i1imcq">cities like Houston</a>, where minority populations are disproportionately affected by the coronavirus, rapid responses are needed to keep communities safe. That’s why Intel is partnering with local companies like Houston-based Medical Informatics Corp (MIC) to provide medical professionals with access to data to deliver remote care and protect essential workers from exposure.</p><p>What does this look like? MIC created its Sickbay platform, which collects patient data through bedside devices in ICUs and other care facilities. This data can then be sent to doctors, nurses, and other caregivers remotely through a PC, tablet, or mobile device to create a virtual intensive care unit. Healthcare workers can then use this software run on Intel architecture to monitor patient analytics at scale remotely.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/997/1*IX7_Nq45nI4DfTConiDzhQ.png" /></figure><p>Combatting the coronavirus isn’t the only obstacle. For large urban centers, creating a safe return to normalcy, however that may look, is just as important.</p><p>In 2019, Intel joined forces with the City of Houston to create smart city solutions through the <a href="https://ionhouston.com/ion-accelerator/">ION Smart and Resilient Cities Accelerator</a>. While this plan was formulated before the COVID-19 pandemic, it can still be used to create a more resilient Houston.</p><p>As the economy slowly reopens, with citizens returning to large gathering places like movie theaters, offices, and schools, <a href="https://www.intel.com/content/www/us/en/internet-of-things/reimagining-life-cities-ebook.html">ensuring a safe return</a> is a top priority. The ION Smart and Resilient Cities Accelerator is partnering with local startups to respond to the pandemic to ensure essential workers like medical professionals, grocery workers, transit workers, and others can return to work safely without putting their health or job on the line. That’s why Houston is now investing in Smart Spaces, where fever checks, contact tracing, and robotics can create a more sustainable city.</p><h4>Autonomous Driving and Transportation Infrastructure</h4><p>While autonomous driving might seem like it belongs in the distant future, solutions are being made to make this an ordinary reality today. Autonomous cars are essential for creating Smart Cities, and at Intel, our technologies are currently powering millions of in-vehicle technologies at this very moment.¹</p><p><a href="https://www.mobileye.com/">Mobileye</a>, an Intel company, is building a robust autonomous vehicle system that can drive solely based on cameras that allows Mobileye systems to pinpoint the critical safety segments for which we truly need redundancy from <a href="https://www.britannica.com/technology/radar">radars </a>and <a href="https://www.britannica.com/technology/lidar">lidars</a>.</p><p>Streetlights might seem like a simple structure, but they can be transformed to connect cities to improve transportation. <a href="https://software.intel.com/content/www/us/en/develop/articles/mindteck-smart-lighting-and-parking.html">Mindteck</a>, an information technology company, created Zigbee, which is a smart lighting solution run on Intel architecture-based IoT gateways. Smart lighting solutions have the ability to enable street lights to be programmed or manually turned on, off, or dimmed from a remote location through Zigbee wireless networks.</p><p><a href="https://www.mindteck.com/MTK_SCS_FIN071118">Mindteck’s smart lighting solution</a> can reduce lighting costs, increase visibility into building operations, and enhance occupant experience. Other smart lighting solutions are being developed to improve transportation, as well. Take <a href="https://www.intel.com/content/dam/www/public/us/en/documents/white-papers/data-drive-quality-smart-city-white-paper.pdf">San Diego, California</a>, for example.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FBWU0SEqEfbo%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DBWU0SEqEfbo&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FBWU0SEqEfbo%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/24e2ebe0b237af4c4f91e9ae0fb0a08e/href">https://medium.com/media/24e2ebe0b237af4c4f91e9ae0fb0a08e/href</a></iframe><p>To improve city living, San Diego placed 3,200 <a href="https://www.sandiego.gov/sites/default/files/fact_sheet_final_smart_street_lights_053119.pdf">CityIQ</a> intelligent nodes, which is a smart lighting solution run on Intel® technology, powered by GE, built by Current, and offered through AT&amp;T. These smart nodes integrate LED lighting with wide-angle cameras and other sensors, preconfigured networking, and local intelligence, and can be installed in any streetlight. Through this citywide IoT network, San Diego reduced 60% of energy use in its streetlights and saved $2.8 million in annual energy costs.</p><p>That’s not the only measure San Diego took to create a Smart City. The San Diego Metropolitan Transit System implemented the Cisco Connect Mass Transit solution, which provides passengers with fast Wi-Fi to receive real-time schedule updates, video surveillance to improve passenger safety, and more energy-efficient routes to save fuel.</p><p>Parking is another factor that leads to traffic congestion and fuel consumption. Through smart parking technology, video analytics can be used in tandem with machine learning to determine vehicle occupancy. <a href="https://software.intel.com/content/www/us/en/develop/articles/mindteck-smart-lighting-and-parking.html">Mindteck</a>, an information technology company, uses Intel architecture to create smart solutions through IoT services. One way Mindteck is helping build Smart Cities is with its smart parking technology.</p><p>Wireless magneto-sensors, ultrasonic sensors, and Wi-Fi cameras can manage on-street, off-street, and multilevel parking by placing parking sensors in parking spaces, which can update routers based on changes in-vehicle presence. This data can then be sent to an individual’s smartphone mobile applications to alert drivers whether a parking space is open or not to reduce time spent searching for one.</p><h4>Sustainability and Conservation</h4><p>Sustainability is one of Intel’s top priorities when it comes to powering Smart Cities. Intel and Bosch are partnering to help cities analyze air pollution data to lower urban power consumption, connect cities to make them smarter, and spread awareness of energy solutions.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FbHzL4KXqFTQ%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DbHzL4KXqFTQ&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FbHzL4KXqFTQ%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/d6f0c66f4f79a628658d988d822968a6/href">https://medium.com/media/d6f0c66f4f79a628658d988d822968a6/href</a></iframe><p>Cities are notoriously known for risky air quality — just take New York City, Los Angeles, and Beijing, for example. To improve people’s lives across the planet, we’ve launched the <a href="https://www.intel.com/content/www/us/en/internet-of-things/bosch-smart-city.html">Intel®-based Bosch Air Quality Micro Climate Monitoring System</a> (MCMS) to help governments, communities, and industries measure, view, and analyze data from numerous spatially spread devices. In turn, these entities will be able to make real-time air quality decisions to make cities smarter, cleaner, and greener for future generations. IoT systems are changing the nature of city living in more ways than one.</p><p>By integrating air quality monitoring systems with other systems like smart traffic networks, smart lighting, and smart parking, city planners can create solutions to ease traffic jams, detect air pollution, and bring positive change to metropolitan centers.</p><h4>The Bottom Line</h4><p>Smart cities are not a thing of the future. The technologies that can make cities smarter are here today.</p><h4>Notices and Disclaimers</h4><p>This article was sponsored by Intel</p><p>Source(s):</p><p>¹ <a href="https://www.mobileye.com/">https://www.mobileye.com/</a></p><p>² <a href="https://venturebeat.com/2018/05/17/intels-mobileye-gets-self-driving-tech-deal-for-8-million-cars/">https://venturebeat.com/2018/05/17/intels-mobileye-gets-self-driving-tech-deal-for-8-million-cars/</a></p><p>© 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.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e664e084d6f0" width="1" height="1" alt=""><hr><p><a href="https://medium.com/intel-tech/what-are-some-of-the-technologies-powering-smart-cities-today-e664e084d6f0">What are Some of the Technologies Powering Smart Cities?</a> was originally published in <a href="https://medium.com/intel-tech">Intel Tech</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[As Networks Get Faster, High-Speed Storage Becomes More Important Than Ever]]></title>
            <link>https://medium.com/intel-tech/as-networks-get-faster-high-speed-storage-becomes-more-important-than-ever-8d502ec8dbbd?source=rss-a5e564ccb3ea------2</link>
            <guid isPermaLink="false">https://medium.com/p/8d502ec8dbbd</guid>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[cloud-computing]]></category>
            <category><![CDATA[data-center]]></category>
            <category><![CDATA[ram-memory]]></category>
            <dc:creator><![CDATA[Intel Tech]]></dc:creator>
            <pubDate>Fri, 04 Jun 2021 11:54:11 GMT</pubDate>
            <atom:updated>2021-06-04T15:26:59.094Z</atom:updated>
            <content:encoded><![CDATA[<h4>Cloud Computing</h4><h4>Part 1 of Intel’s Cloud Service Providers Solutions Series</h4><p>Author(s): Jacek Wysoczynski, Intel Senior Manager of Product Planning and Andrew Ruffin, Intel Strategy &amp; Business Development Manager</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/588/1*5ImRe6xwFxCjx3YO2Nc_hg.png" /><figcaption>Intel Photo of Blue lit Data Center</figcaption></figure><p>The demands on Cloud Service Providers (CSPs) are increasing rapidly. They need to provide more storage and more compute power, all while keeping network performance up and reducing downtime. As CSPs upgrade their networks to 100 Gigabit Ethernet (GbE), the need to protect incoming data as it’s being written to bulk storage is increasingly more important.</p><p>Currently, CSPs typically use nonvolatile DIMMs (NVDIMMs) or over-provisioned NAND solid state drives (SSDs) as write or power loss imminent (PLI) buffers. However, both of these hardware choices have limitations.</p><p>Intel Optane SSDs allow data center operators to gain the benefits of near-NVDIMM speed with the ease of upgrade and repairs of NAND SSDs, all in one system. And, as network speeds increase, Intel Optane SSDs improve the situation further, by helping to reduce overall operating costs and simplifying data center maintenance.</p><p>This blog is part of a two-part series about improving Data Center operations with Intel products. In the first part of this series we will review the advantages and disadvantages data center operators had to deal with in the past and in part 2 of the series we will look at the Intel Optane SSD options that address these challenges.</p><p>NVDIMMs are a great technology, combining the speed of Dynamic RAM (DRAM) with the non-volatile memory protection of NAND SSDs. Using NVDIMMs, servers experiencing power loss can have their data saved to the NAND flash on the NVDIMM, protecting that data from loss. Once power is restored to the server, the protected data can then be transferred to bulk storage. With that said, there are questions that should be asked with regards to which technologies should be used in a data center? Here are 5 questions we should ask ourselves when choosing memory and storage solutions for a data center:</p><p><strong>1.</strong> <strong>Is NVDIMM capacity inadequate based on rising throughput?</strong></p><p>The amount of data coming through network connections is increasing. 100 GbE network connections are becoming more and more common in data centers, and faster connections could be here soon. As the amount of information coming to a server increases, the current capacity of NVDIMMs will no longer meet write buffer sizing needs. Adding in additional NVDIMM capacity to a server is limited by multiple constraints, including the capacity of each memory module, the number of available slots for NVDIMM in the server, and the cost of the NVDIMM. For example, 32GB of DIMM can handle 2.5 seconds of data streaming through a 100GbE connection.</p><p>This data coming from the NVDIMM then needs to be de-staged to the bulk storage. If the server’s bulk storage is not able to keep up with the NVDIMMs de-staging, the system must reject new write operations. The data write requests must be repeated, which can cause network congestion, as more data can still come in while the prior data is still being processed.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/624/1*bmIfLmGGg7lHG7sJZLlI4Q.png" /><figcaption>Intel Photo of Data Center Network Cabling</figcaption></figure><p><strong>2.</strong> <strong>Do NVDIMMs increase system downtime for upgrades and repairs?</strong></p><p>NVDIMM is much more difficult to replace in servers than technology such as SSDs. SSDs are hot-swappable; that is, they can be added or removed from a computer or server without the need to shut down. When attached to servers, SSDs are typically also front-serviceable, where the bay holding the SSD can be accessed while the server is still on the rack. NVDIMMs, on the other hand, are neither hot-swappable, nor front-serviceable. Adding new or replacing old NVDIMMs requires the servers to be shut down and removed from the rack. This increases the time that servers are offline and unavailable for use. This increase in downtime and additional complexity for upgrades and repairs adds to the overall cost of maintaining data centers.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/624/1*v43ooiNjvxX5dMP15ADuEA.png" /><figcaption>Intel Photo of Data Center Rack Review</figcaption></figure><p><strong>3. Are NVDIMMs less cost effective than SSDs?</strong></p><p>In addition to the higher cost of service for NVDIMMs vs. SSDs, NVDIMMs themselves are more <strong>expensive </strong>than an SSD in terms of price per GB². NVDIMMs combine DIMM RAM with a battery or supercapacitor to protect data in the event of a loss of power, which increases the price further. While comparing the price of DIMM vs. SSD is not perfect, for data centers trying to control costs, using all NVDIMM for their PLI buffers increases the overall price of buying and maintaining servers.</p><p><strong>4. What are the limitations associated with NAND SSDs for Cloud Storage?</strong></p><p>SSDs have been in use for a long time. Data Center NAND SSDs can act as an ideal write buffer between DRAM and bulk storage. SSDs are typically much cheaper than NVDIMMs, and they are easier to upgrade and replace. Much like NVDIMMs, however, there are drawbacks to using them in data centers.</p><p>The latency on NAND SSDs is much higher compared to RAM and NVDIMM. Latency is the delay between when an action is requested and when it’s completed. This means a lower latency is more desirable than a higher one when processing data. RAM, including NVDIMM, has significantly lower latency compared to SSDs. Latency on RAM is typically counted in nanoseconds, while for SSDs it’s counted in microseconds. The latency of the fastest NAND SSDs is around 1000 times greater than that of RAM. When dealing with large amounts of incoming information, such as in a data center, this high latency can prevent increases in network speeds from being properly utilized.</p><p>SSDs, like most hardware, have a limited lifespan before repairs and/or replacements are necessary. With SSDs, the lifespan is controlled by the number of writes each data block can handle before data can no longer be safely stored there, estimated in Drive Writes Per Day (DWPD), or the number of times per day the entire SSD can be rewritten over the lifespan of the drive. A NAND SSD might, for example, be rated for three DWPD over the course of a five-year period. In a busy data center, if the data size is bigger than this number, then the SSD would need to be replaced sooner than this five-year lifespan.</p><p>RAM and NVDIMM has a much longer lifespan in comparison, as data can be written and rewritten to it without significantly degrading its function. So, while RAM is most often replaced due to a need for more speed or because of hardware failure, SSDs may be switched out because data can no longer be written to them. These SSDs must then be replaced to keep the server operating.</p><p><strong>5. How does NAND SSD Data Write Protection and Overprovisioning work?</strong></p><p>NAND SSDs are broken up into blocks and pages. Each page is 4KB, and each block typically contains 128 pages. Data is written one page at a time, but once data must be rewritten, it cannot be done on a page that already has data; that page must first be erased. However, individual pages cannot be erased, only entire blocks. Instead, changes made are written to empty pages, and the original pages are marked as invalid. Later, when space is needed, an entire block is rewritten to a new block, minus any pages that are marked as invalid, and the old block is then erased.</p><p>This process, known as garbage collection, is one of the reasons why NAND SSDs have higher latency and inconsistent performance versus RAM, as data can be moved multiple times on a SSD to accommodate these writes and rewrites. More data must be moved in order to allow for blocks to be erased and reused. It also requires the SSD to have additional space just to allow these extra write operations. This is commonly referred to as overprovisioning, and all NAND SSDs have about 7% of their total storage dedicated to this1. Users can increase overprovisioning, setting aside more of the SSD for write operations. These changes increase the SSD’s write speeds, as there is more space available to write incoming data. It can also increase the SSD’s lifespan by improving the DWPD, although this requires having more drive space, as more is set aside for write functions.</p><p>When more of the NAND SSD is set for overprovisioning, it reduces the amount of space available for usable storage. For example, if a 500GB SSD is set up with 50% overprovisioning, its actual storage space would be reduced to 250GB. This would mean that if the system needed at least 500GB in storage, then it would either require two SSDs set up this way, or one SSD with 1TB of storage. While overprovisioned SSDs would be faster than the non-overprovisioned ones, they would still not be as fast as NVDIMM, and the reduction in storage space would increase the costs.</p><p>In the next part of this two-part Cloud Storage solution series, we will review Intel’s solutions for the Data Center and how these solutions address the issues raised in this blog.</p><h4>Notices and Disclaimers</h4><p>Source(s):</p><p>¹ <a href="https://www.seagate.com/tech-insights/ssd-over-provisioning-benefits-master-ti/">Seagate Over Provisioning and Its Benefits-Seagate.com</a></p><p>² <a href="https://searchstorage.techtarget.com/answer/NVDIMM-vs-memory-channel-flash-storage">NVDIMM vs. memory channel flash storage by Marc Staimer of Dragon Slayer Consulting</a></p><p>Performance varies by use, configuration and other factors. Learn more at <a href="http://www.intel.com/PerformanceIndex">www.Intel.com/PerformanceIndex</a>.</p><p>© 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.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8d502ec8dbbd" width="1" height="1" alt=""><hr><p><a href="https://medium.com/intel-tech/as-networks-get-faster-high-speed-storage-becomes-more-important-than-ever-8d502ec8dbbd">As Networks Get Faster, High-Speed Storage Becomes More Important Than Ever</a> was originally published in <a href="https://medium.com/intel-tech">Intel Tech</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How We at Intel Optimized for TensorFlow 2.5…]]></title>
            <link>https://medium.com/intel-tech/how-we-at-intel-optimized-for-tensorflow-2-5-1ce908c52211?source=rss-a5e564ccb3ea------2</link>
            <guid isPermaLink="false">https://medium.com/p/1ce908c52211</guid>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[deep-learning]]></category>
            <category><![CDATA[tensorflow]]></category>
            <category><![CDATA[data-science]]></category>
            <dc:creator><![CDATA[Intel Tech]]></dc:creator>
            <pubDate>Mon, 24 May 2021 18:07:24 GMT</pubDate>
            <atom:updated>2021-05-24T20:33:54.102Z</atom:updated>
            <content:encoded><![CDATA[<h4>Machine Learning</h4><h4>A principal engineer’s approach to optimizing TensorFlow performance on Intel® Xeon® processors</h4><p>Author: Ramesh AG, Intel Principal Engineer</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hrHQPbGeHGSI0Ymci00bUg.png" /><figcaption>Team of Computer Engineers Work on Machine Learning Neural Network Technology Development, Intel photos</figcaption></figure><p>I am a principal engineer at Intel, where I have worked for more than 25 years, largely on software optimizations. For the last four years my primary responsibility has been working with Google engineers to optimize the TensorFlow artificial intelligence (AI) framework so it runs as fast as possible on Intel Xeon CPUs.</p><p>When we started this effort four years ago, much of the talk about machine learning (ML) and in particular deep learning focused on GPUs. Mistakenly, Intel Xeon CPUs were often not perceived as an alternative for deep learning. Frankly, one reason was that the most popular framework for machine learning, TensorFlow, was not running as well as it should on our Intel Xeon CPUs.</p><p>Over the last three years, working with Google, our group at Intel has performed many optimizations to TensorFlow, and now those improvements are part of the latest 2.5 release of TensorFlow. You can read more about these optimizations and what we did in the Intel Analytics Software blog, <a href="https://medium.com/intel-analytics-software/leverage-intel-deep-learning-optimizations-in-tensorflow-129faa80ee07"><em>Leverage Intel Deep Learning Optimizations in TensorFlow</em></a>.</p><p>Developers now recognize that machine learning performs very well on a multi-core CPU. Since there are millions of Intel Xeon CPUs in operation¹, it is a good thing for AI in general that these algorithms run well on a multi-core CPU, so people don’t have to think that they have to go out and buy an expensive GPU. We are helping make that possible by making TensorFlow and ML applications run faster on Intel hardware, which is already widely available at data centers, for students in colleges, and many other places.</p><p>The second thing I would like to emphasize is that Intel has worked on new instructions for AI processing, in our Intel® Deep Learning Boost software. Intel DL Boost has specific instructions which accelerate workloads that use smaller data types than the typical ML data type of 32 bits. What the AI industry has seen is that machine learning in general does not require 32-bit floating-point numbers, and can tolerate smaller data types, for example, 8-bit and 16-bit data types. We have specific instructions for these matrix multiplications, using 8-bit data types. These smaller data types are becoming popular, and Intel DL Boost addresses that, among other topics.</p><p>Applications run much faster on Intel Xeon processors by using the special optimizations that we have added to TensorFlow. We have proven that by conducting benchmark testing and publishing the results for Intel’s Xeon processors on the <a href="https://medium.com/intel-analytics-software/leverage-intel-deep-learning-optimizations-in-tensorflow-129faa80ee07"><em>Leverage Intel Deep Learning Optimizations in TensorFlow</em></a>, which over the past two years are the most widely deployed processors in the cloud. Depending on the SKU, these Xeon processors can have up to 40 cores per socket, with two sockets per system for up to 80 cores per system. One important aspect of our optimizations is to make sure that the workloads are balanced for the number of available cores.</p><p>That was an important part of our optimization effort, to make sure the machine learning algorithms are properly partitioned over all the cores that are available. The matrix math operations can be broken down into smaller pieces so that each one can be done on a separate core individually, and then combined to get the results.</p><p>These hardware and software optimizations are continuing, with our latest Intel Xeon scalable processors as a prime example on the hardware side and our TensorFlow work with Google, and Intel DL Boost, as good examples on the software side. This work is necessary because the AI models are becoming much more complex and Intel is continuing to innovate to address these and future complexities.</p><h4>Notices and Disclaimers</h4><p>¹Sources:</p><ul><li><a href="https://www.storagereview.com/news/intel-3rd-gen-xeon-scalable-processors-ice-lake">Intel 3rd Gen Xeon Scalable Processors — Ice Lake</a></li><li><a href="https://newsroom.intel.com/news/xeon-scalable-5g-network-portfolio-launch/#gs.19z90n">Intel Reinforces Data Center Leadership with New 2nd Gen Intel Xeon Scalable Processors</a></li></ul><p>Performance varies by use, configuration and other factors. Learn more at <a href="https://edc.intel.com/content/www/us/en/products/performance/benchmarks/overview/">www.Intel.com/PerformanceIndex​.com</a>.</p><p>Intel technologies may require enabled hardware, software or service activation. No product or component can be absolutely secure.<br>© 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.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1ce908c52211" width="1" height="1" alt=""><hr><p><a href="https://medium.com/intel-tech/how-we-at-intel-optimized-for-tensorflow-2-5-1ce908c52211">How We at Intel Optimized for TensorFlow 2.5…</a> was originally published in <a href="https://medium.com/intel-tech">Intel Tech</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Leverage Intel Deep Learning Optimizations in TensorFlow]]></title>
            <link>https://medium.com/intel-analytics-software/leverage-intel-deep-learning-optimizations-in-tensorflow-129faa80ee07?source=rss-a5e564ccb3ea------2</link>
            <guid isPermaLink="false">https://medium.com/p/129faa80ee07</guid>
            <category><![CDATA[tensorflow]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[deep-learning]]></category>
            <category><![CDATA[neural-networks]]></category>
            <dc:creator><![CDATA[Intel Tech]]></dc:creator>
            <pubDate>Fri, 14 May 2021 22:42:40 GMT</pubDate>
            <atom:updated>2021-05-14T22:43:15.556Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*KJ8YJD3qKT3Wqyukp17i4Q.png" /></figure><h4>Artificial Intelligence</h4><h4>Set a Single Environment Variable to Get up to 3x Performance Boost</h4><p>TensorFlow is a widely-used deep learning (DL) framework. Intel has been collaborating with Google to optimize its performance on Intel Xeon processor-based platforms using Intel oneAPI Deep Neural Network (oneDNN), an open-source, cross-platform performance library for DL applications. TensorFlow optimizations are enabled via oneDNN to accelerate key performance-intensive operations such as convolution, matrix multiplication, and batch normalization.</p><p>We are happy to announce that the oneDNN optimizations are now available in the official TensorFlow release, enabling developers to seamlessly benefit from the Intel optimizations. Additional TensorFlow-based applications, including TensorFlow Extended, TensorFlow Hub, and TensorFlow Serving will also include the oneDNN optimizations.</p><h3><strong>Enabling oneDNN Optimizations in TensorFlow 2.5</strong></h3><ol><li>Install the latest TensorFlow pip package, i.e.: pip install tensorflow</li><li>By default, the oneDNN optimizations will be turned off. To enable them, set the environment variable TF_ENABLE_ONEDNN_OPTS. On Linux systems, for example: export TF_ENABLE_ONEDNN_OPTS=1</li><li>Run your TensorFlow application.</li></ol><h3><strong>Performance Benefits of TensorFlow 2.5 with oneDNN Optimizations</strong></h3><p>We benchmarked several popular TensorFlow models on DL inference and training, comparing results with oneDNN optimizations enabled on a 2nd Generation Intel Xeon Scalar Processor platform.</p><p>Inference was benchmarked using four cores on a single-socket for latency measurements with all 28 cores for throughput tests. Figures 1 and 2 show the relative performance improvement for inference across a range of the models. For offline throughput measurements (using large batches), performance improvements up to 3x are possible (Figure 1). For real-time server inference (batch size = 1), the oneDNN-enabled TensorFlow took 29% to 77% the time of the unoptimized version for 10 out of 11 models (Figure 2).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/964/1*98x1AG6IBqXqOeNEp4NhnA.png" /><figcaption>Figure 1. Inference throughput improvements</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/966/1*XToWHj1twmMyEQSDEdMZ8w.png" /><figcaption>Figure 2. Inference latency improvements</figcaption></figure><p>For training, up to 2.4x performance gains were observed across several popular models (Figure 3). Gains were also observed with previous TensorFlow 1.x graph models and the newer TensorFlow 2.x eager execution-based models.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/913/1*wCVW1vMpY3nCXh-yBiqTjg.png" /><figcaption>Figure 3. Training performance improvements</figcaption></figure><p>You can reproduce these benchmarks by getting the same models from <a href="https://github.com/IntelAI/models">Model Zoo for Intel Architecture</a>:</p><pre>git clone <a href="https://github.com/IntelAI/models.git">https://github.com/IntelAI/models.git</a></pre><p>The README.md files contain instructions to perform model training and inference. For example, the instructions for the inceptionv3 model are in models/benchmarks/image_recognition/tensorflow/inceptionv3/README.md.</p><p>Based on the benchmarking and results above, we encourage data scientists and developers to download the latest official TensorFlow release and enable the oneDNN optimizations to get immediate performance improvements on Intel Xeon processor-based platforms.</p><h3><strong>Low Precision Data Type</strong></h3><p>oneDNN also enables the int8 and <a href="https://cloud.google.com/blog/products/ai-machine-learning/bfloat16-the-secret-to-high-performance-on-cloud-tpus">bfloat16</a> data types to improve compute-intensive training and inference performance on the latest 2nd- and 3rd-Generation Intel Xeon Scalable processors. These optimizations can improve model execution time by up to <a href="https://www.intel.com/content/www/us/en/artificial-intelligence/posts/deep-learning-performance-boost-by-intel-vnni.html">4x for int8</a> and <a href="https://blog.tensorflow.org/2020/06/accelerating-ai-performance-on-3rd-gen-processors-with-tensorflow-bfloat16.html">2x for bfloat16</a>. The int8 data type is not currently supported in the official TensorFlow 2.5 release, but this limitation will be addressed in a later version. In the meantime, to use the int8 data type, you can download the <a href="https://software.intel.com/content/www/us/en/develop/articles/intel-optimization-for-tensorflow-installation-guide.html">Intel Optimization for TensorFlow.</a></p><h3><strong>TensorFlow Resources and Support</strong></h3><p>TensorFlow 2.5 can be found here:</p><ul><li><a href="https://github.com/tensorflow/tensorflow/releases">Official TensorFlow release</a></li><li><a href="https://pypi.org/project/tensorflow/">Pip packages</a></li><li><a href="https://hub.docker.com/r/tensorflow/tensorflow/">Docker</a> images</li></ul><p>For help with technical questions, visit the following communities and forums to find answers and get support:</p><ul><li><a href="https://github.com/tensorflow/tensorflow/issues">TensorFlow GitHub issues</a></li><li><a href="https://software.intel.com/en-us/forums/intel-optimized-ai-frameworks">Intel Optimized AI Frameworks</a></li></ul><h3><strong>Benchmarking System Configuration</strong></h3><p>Two-socket Intel Xeon Platinum 8280L Processor, 28 cores, HT On, Turbo On, total memory 256GB.</p><p>System BIOS: SE5C620.86B.02.01.0012.070720200218.</p><p>TensorFlow version: <a href="https://github.com/tensorflow/tensorflow.git">https://github.com/tensorflow/tensorflow.git</a> (2.5RC3)</p><p>Compiler and libraries: gcc 7.5.0, oneDNN v2.2.0</p><p>Datatype: FP32</p><p>Data collection date: May 9, 2021</p><h3>Contributors</h3><p>Mahmoud Abuzaina, Deep Learning Software Engineer; Ramesh AG, Principal Engineer; Jason Chow, Marketing Manager; Xiaoming Cui, Deep Learning Software Engineer; Rama Ketineni, Deep Learning Software Engineer; and Guozhong Zhuang, Deep Learning Software Engineer; Intel Corporation.</p><h4><strong><em>Notices and Disclaimers</em></strong></h4><p>Performance results are based on testing as of dates shown in configurations and may not reflect all publicly available options. Learn more at <a href="https://edc.intel.com/content/www/us/en/products/performance/benchmarks/overview/">www.Intel.com/PerformanceIndex​</a>.</p><p>Intel technologies may require enabled hardware, software or service activation. No product or component can be absolutely secure.</p><p>© 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.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=129faa80ee07" width="1" height="1" alt=""><hr><p><a href="https://medium.com/intel-analytics-software/leverage-intel-deep-learning-optimizations-in-tensorflow-129faa80ee07">Leverage Intel Deep Learning Optimizations in TensorFlow</a> was originally published in <a href="https://medium.com/intel-analytics-software">Intel Analytics Software</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Integrated Photonics Set to Light Up the Data Center]]></title>
            <link>https://medium.com/intel-tech/integrated-photonics-set-to-light-up-the-data-center-eb7962e51e92?source=rss-a5e564ccb3ea------2</link>
            <guid isPermaLink="false">https://medium.com/p/eb7962e51e92</guid>
            <category><![CDATA[onezero]]></category>
            <category><![CDATA[computer-science]]></category>
            <category><![CDATA[fiber-optics]]></category>
            <category><![CDATA[science]]></category>
            <category><![CDATA[technology]]></category>
            <dc:creator><![CDATA[Intel Tech]]></dc:creator>
            <pubDate>Tue, 11 May 2021 00:37:26 GMT</pubDate>
            <atom:updated>2021-05-14T22:11:33.846Z</atom:updated>
            <content:encoded><![CDATA[<h4>Technology</h4><h4>Author: Jeff W. Hockert, Senior Marketing Manager at Intel</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jXG3bVqzKs1TcPu07qeP9Q.png" /><figcaption>Intel image of Intel Integrated Photonics Chip concept</figcaption></figure><p>As electrical interconnects face performance barriers and data centers confront a power wall, Intel is advancing its vision of Integrated Photonics. By transmitting data optically, at lower latencies, higher bandwidth, and much lower power than is possible with electrical connections, Intel believes Integrated Photonics will fundamentally change data center architectures.</p><p>Intel’s <a href="https://www.intel.com/content/www/us/en/architecture-and-technology/silicon-photonics/silicon-photonics-overview.html"><strong>Silicon Photonics</strong></a> vision began in 2004 with a goal of pulling together the optical and electronic functions into one chip. A lot of progress has been made since then. For example, there are now over 4 million Intel 100G optical transceivers being used for rack-to-rack connectivity — connecting row switches and director-class switches throughout data centers.</p><p>Today, optical interconnects over fiber dominate long-distance interconnects, while electrical dominates short distances. Our goal is to advance optical performance and replace electrical by bringing optical I/O directly into servers for board-to-board and package-to-package connectivity.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/624/1*G_R7o-e2Gwv7QySwZcMQAQ.png" /><figcaption>Figure 1 Moving to Interconnects slide source Intel</figcaption></figure><p>And the need is there: current I/O performance of electrical interconnects are running into practical performance limits. Equally important, data centers are faced with power-consumption challenges. Bandwidth demand for compute approximately doubles every two years¹. However, electrical performance-scaling is not keeping pace with bandwidth demand. Moreover, electrical I/O energy efficiency has slowed recently, resulting in an approaching power wall. Specifically, the I/O power is trending to be larger than the total available socket power, leaving nothing for compute.</p><p>In traditional server systems, scaling electrical interconnect requires higher data rates and adding more I/O pins, which increases the package size to meet bandwidth demand. However, this doesn’t solve the power wall. Without fundamental breakthroughs, these two trends will not change over time. Optical I/O has the potential to be more efficient than electrical I/O while simultaneously providing &gt;1000 times more reach². By tightly integrating optical into a server package using silicon photonics, we can gain three benefits in a smaller footprint: (1) lower power (2) higher bandwidth and (3) reduced pin count.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tALSwYw2SP6Zk4V-JM--dw.png" /><figcaption>Intel Integrated Photonics image source Intel.com</figcaption></figure><p>We call this research vision Integrated Photonics and believe it will fundamentally change our data center network architectures. It will free data to move around the data center much more efficiently — at both lower power and lower latency. Our research goals are 1Tb/s per fiber at @ 1pJ/b with up to a 1km reach. By reducing the optical I/O power below electrical I/O and miniaturizing the footprint of our silicon photonics devices to drive down cost, we are confident we can address the approaching I/O power wall.</p><p>Let’s drill down and see what technologies are needed to achieve Intel’s vision of connecting data centers with light. At <a href="https://newsroom.intel.com/press-kits/intel-labs-day-2020/">Intel Labs Day 2020</a>, held on Dec. 3rd, Intel showed photonics ICs with fully integrated photonics functions, including the laser, amplifier, and ring modulators/detectors. We showed breakthrough advances in (1) micro-ring modulators, (2) an all-silicon photodetector, (3) optical amplifiers integrated with the lasers and (4) CMOS circuit integration.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ot4st80GdyZ2h_sX3h9B-g.png" /><figcaption>Intel Silicon Photonics building blocks image source Intel.com</figcaption></figure><p>We have developed micro-ring modulators which are a thousand times smaller than conventional silicon modulators. The micro rings serve as a light modulator, and as a silicon photodetector. Until now, the industry believed that silicon had virtually no light detection capability in the 1.3–1.6um wavelength range, seen as a fundamental limitation for silicon photonics technology. That was proved wrong. On the receiver side, we have been working on all silicon ring-based photodetectors which combine wavelength selection and photodetection functions in a single device. Earlier this year we published light detection capability with an all-silicon photodetector, and demonstrated this photodetector operating at a data rate of 112Gb/s.</p><p>These ring photodetectors can be assembled with CMOS trans-impedance amplifiers to build low-cost optical receivers. A major advantage of this approach is processing and material cost reduction.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ptUNghNS3M8IuKRE1Bu4Jg.png" /><figcaption>Intel Integrated Photonics building blocks image — source Intel.com</figcaption></figure><p>We have integrated multi-wavelength lasers. We showed a 4-wavelength laser which uses a single cavity, making it a more compact laser with uniform channel spacing. Semiconductor optical amplifiers are another indispensable technology. It is challenging to generate all the required light power from a laser. Fortunately, Intel’s optical amplifiers can be more power efficient, and it is important to note that they are made from the same material used for the integrated laser. Without the integrated laser, these amplifiers are not possible.</p><p>Finally, the co-integration of CMOS circuitry and silicon photonics brings all these technologies together. Early this year we described our ability to 3D stack the CMOS electronic IC with our photonics IC using copper pillars. The CMOS electronic IC includes interface circuits to transmit, receive, and thermally control the ring modulator to track temperature variations.</p><p>By directly stacking the two ICs, one on top of the other, we demonstrated the real potential of micro-rings and CMOS circuitry being integrated together. The die-to-die integration results in a 3D-integrated 100G silicon photonic transmitter, capable of signaling at 112 Gbps.</p><p>With excellent package integration technology developed by Intel’s packaging group, we have a path to integrate the most advanced CMOS technology and silicon photonics for bringing products with unmatched interconnect performance to our customers. Intel is unique in its ability to deliver integrated multi-wavelength lasers, semi-conductor optical amplifiers, all-silicon photodetectors, and micro-ring modulators on a single silicon technology platform, along with the technology to tightly integrate them with CMOS silicon.</p><p>Integrating photonics onto a package for servers is not a new idea. The key difference is that we now have all the required technologies on a single silicon photonics platform to make this vison a reality. With our micro-ring modulators being reduced to just microns in size, we can integrate 100s of optical I/O in a small footprint. We believe these types of optical links demonstrated at <a href="https://newsroom.intel.com/press-kits/intel-labs-day-2020/">Intel Labs Day 2020</a> can be scaled to enable bandwidths of 10s to 100s of terabits per second, and will increase optical I/O volumes from millions today to billions of units, driving down manufacturing costs.</p><p><strong>About the Author</strong></p><p><em>Jeff Hockert is a senior marketing manager in the Technology Leadership Marketing (TLM) team at Intel.</em></p><p><strong><em>Notices and Disclaimers</em></strong></p><p>¹ Source(s):</p><ul><li><a href="https://www.itpro.co.uk/604527/study-predicts-serious-bandwidth-shortages">ITPRO. — “Study predicts serious bandwidth shortages” by Miya Knights ‘08</a></li><li><a href="https://spectrum.ieee.org/telecom/wireless/edholms-law-of-bandwidth">Edholm’s law</a></li></ul><p>² Source: f<a href="https://fiberopticalworld.com/copper-vs-fiber-difference-which-one-is-the-best-choice-draw-backs/">iber optical world, <strong>Copper Vs Fiber | Difference | Which One is the best choice | Draw backs, </strong>by Saeed ‘20</a></p><p>Performance varies by use, configuration and other factors. Learn more at <a href="https://edc.intel.com/content/www/us/en/products/performance/benchmarks/overview/">www.Intel.com/PerformanceIndex​.com</a></p><p>Intel technologies may require enabled hardware, software or service activation. No product or component can be absolutely secure. <br> © 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.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=eb7962e51e92" width="1" height="1" alt=""><hr><p><a href="https://medium.com/intel-tech/integrated-photonics-set-to-light-up-the-data-center-eb7962e51e92">Integrated Photonics Set to Light Up the Data Center</a> was originally published in <a href="https://medium.com/intel-tech">Intel Tech</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Security Built on a Foundation of Trust]]></title>
            <link>https://medium.com/intel-tech/security-built-on-a-foundation-of-trust-1fa1dbb74cbc?source=rss-a5e564ccb3ea------2</link>
            <guid isPermaLink="false">https://medium.com/p/1fa1dbb74cbc</guid>
            <category><![CDATA[security]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <dc:creator><![CDATA[Intel Tech]]></dc:creator>
            <pubDate>Fri, 16 Apr 2021 23:16:06 GMT</pubDate>
            <atom:updated>2021-05-11T19:40:10.139Z</atom:updated>
            <content:encoded><![CDATA[<h4><strong>Cybersecurity</strong></h4><h4><strong>Author: Brian Richardson, Security Pillar Lead, Intel Corporation</strong></h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/624/1*qtw566ixKQTXXvYlb4HPng.png" /></figure><p><strong>Intel Platform Security is No Accident</strong></p><p>As the 21st century moves along, the demand for accurate cross-platform data exchange continues to increase every day. New technologies are developed, new applications are implemented, and exciting technological avenues are discovered and explored at breakneck speed. Devices and applications that would have been considered sci-fi just a few years ago are finding their way into our pockets and onto our desktops, and public expectations demand that they all work together seamlessly.</p><p>Unfortunately, the connections between these technologies are a tempting target for attackers looking for new and creative ways to undermine the system and gain access to the data — either for notoriety or personal gain. They know that if they nibble around the edges of any system long enough, they will eventually find a vulnerability that can be exploited.</p><p><strong>Top-to-Bottom Focus on Security</strong></p><p>At Intel, we understand that this constant pressure to perform and the equal pressure to protect requires strong, consistent application of sound security strategies. We maintain a steady focus on the potential security risks at every level of our product development and manufacturing processes, beginning at the drawing board, and continuing until long after product delivery. This is accomplished in two main forms:</p><ol><li>The Intel Security Development Lifecycle (SDL) is a comprehensive set of practices we apply to our hardware, software, and firmware to integrate security principles and privacy tenants at every step of product development, from concept through end of servicing.</li><li>Threat discovery and incident response, including dedicated offensive research of our products and emerging security threats, as well identifying vulnerabilities, mitigations, coordinated disclosure to the industry.</li></ol><p><strong>Silicon-Enabled Security</strong></p><p>Critical security features are built into every chip and are present at every level — including the silicon itself. We innovate to protect data and privacy, centered around three key priorities:</p><ul><li>Foundational Security: critical protection to help verify trustworthiness of devices and data.</li><li>Workload and Data Protection: trusted execution for hardware-isolated data protection.</li><li>Software Reliability: platforms that help protect against a range of cybersecurity threats.</li></ul><p>Intel’s hardware components provide the backbone of a strong security foundation to build solid application platforms. The processor logic adds another layer of security, and the resulting product architecture within each chip can be used to support virtually any application.</p><p>With ongoing investments in security research and processes that span the product lifecycle, security doesn’t end at product launch. Instead, security updates are made as needed until end of servicing to continually improve product resilience.</p><p>Providing every legitimate workload with a trusted execution environment for hardware-isolated protection of data in use, scaled to fit workloads of varying sizes. These security technologies extend to help protect virtual machines and operating systems against a variety of targeted attacks.</p><p>Software vulnerabilities represent the biggest opportunity for attackers, who patiently search for any potentially unsecured opening or unintended weakness. Between internal and external security research along with robust incident response, Intel seeks out potential vulnerabilities and works to mitigate them before they can be exploited. Simultaneously, we look for ways to architect security capabilities into the silicon to improve resilience for common software issues.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/624/1*Gr1b28NONEaO55BMTpnj_A.png" /><figcaption>Bug Bounty Program 2020 Overview¹</figcaption></figure><p><strong>It Takes a Village (or an ecosystem)</strong></p><p>We recognize that extensive testing and real-world application are necessary to improving security. So, we rely on the relationships with our developer ecosystem, including external security researchers and partners. These developers use our products to build marketplace solutions and identify potential vulnerabilities.</p><p>We appreciate every opportunity to work with them and create security solutions that work on every level. In addition, we work with programmers from all corners of the world as part of our Bug Bounty Program. This program promotes research and coordinated disclosure of potential vulnerabilities by ethical hackers and security researchers.</p><p>Together, the people and processes employed by Intel ensure a layered security approach that powers devices you can count on. For more information on the people, the processes and the products that make Intel special, please read Martin Dixon’s ‘<a href="https://www.intel.com/content/www/us/en/newsroom/opinion/people-processes-products-define-security-strategy.html">People, Processes, Products Define Intel’s Security Strategy</a>’ article available in the Intel Newsroom.</p><p><strong>About the Author</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/375/1*NujmYjx4pH_GLdkN9Fv7Ug.png" /><figcaption>Brian Richardson</figcaption></figure><p><em>Brian Richardson is the Security Pillar Technology Marketing Lead for Intel Technology Leadership Marketing (TLM).</em> <em>Prior to this role, Brian focused on the firmware that quietly boots billions of computers. He focused on implementation of the Unified Extensible Firmware Interface (UEFI), migrating the industry from 16-bit PC BIOS, and promoting standards-based firmware development (tianocore.org). Brian has presented at technical conferences including Intel Developer Forum, Bsides, LinuxCon, FOSDEM, and UEFI Plugfest.</em> <em>Brian incorporates his interest in photography, video production, and event hosting at Dragon Con into his work in technical marketing. He also incorporates his interest in martial arts into his farm improvement projects by occasionally kicking things that don’t work properly.</em></p><p>(1) <a href="https://www.linkedin.com/in/richardsonbrian/%22%20/t%20%22_blank">Brian Richardson | LinkedIn</a></p><p><em>(2)</em> <a href="https://twitter.com/Intel_Brian%22%20/t%20%22_blank">Brian Richardson (@Intel_Brian) / Twitter</a></p><p><strong><em>Notices &amp; Disclaimers</em></strong></p><p>¹ Source:<a href="https://www.intel.com/content/dam/www/public/us/en/security-advisory/documents/intel-2020-product-security-report.pdf%22%20/t%20%22_blank"> Intel-2020-product-security-report</a></p><p>Performance varies by use, configuration and other factors. Learn more at <a href="http://www.intel.com/PerformanceIndex%22%20/t%20%22_blank">www.Intel.com/PerformanceIndex</a>​.</p><p>Performance results are based on testing as of dates shown in configurations and may not reflect all publicly available ​updates. See sources cited for configuration details. No product or component can be absolutely secure.</p><p>© 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.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1fa1dbb74cbc" width="1" height="1" alt=""><hr><p><a href="https://medium.com/intel-tech/security-built-on-a-foundation-of-trust-1fa1dbb74cbc">Security Built on a Foundation of Trust</a> was originally published in <a href="https://medium.com/intel-tech">Intel Tech</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Ice Lake Advances Confidential Computing with Intel® SGX, Total Memory Encryption]]></title>
            <link>https://medium.com/intel-tech/ice-lake-advances-confidential-computing-with-intel-sgx-total-memory-encryption-df326f5e9800?source=rss-a5e564ccb3ea------2</link>
            <guid isPermaLink="false">https://medium.com/p/df326f5e9800</guid>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[hacker]]></category>
            <category><![CDATA[cyberattack]]></category>
            <category><![CDATA[technews]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <dc:creator><![CDATA[Intel Tech]]></dc:creator>
            <pubDate>Thu, 15 Apr 2021 16:48:04 GMT</pubDate>
            <atom:updated>2021-04-15T16:48:04.879Z</atom:updated>
            <content:encoded><![CDATA[<h3><strong>Ice Lake Advances Confidential Computing with Intel SGX, Total Memory Encryption</strong></h3><p><strong>Author: Lisa Spelman, Intel Corporate Vice President and General Manager of the Xeon and Memory Group</strong></p><p>As more businesses and ecommerce move online, cyber criminals are developing new ways to hack the sensitive information of individuals and organizations. In fact, within the first half of 2020, over 36 billion data records were exposed¹. Without action from the ecosystem, this trend will only continue to grow.</p><p>Intel has been a leader in establishing the confidential computing arena -helping customers process and consolidate data with enhanced security and without sacrificing performance or cost. With our latest 3rd Gen Intel Xeon Scalable Processors, we are bringing<a href="http://intel.com/sgx"> Intel Software Guard Extensions </a>(Intel SGX) to our mainstream 2-socket server processors, along with new data protection technologies including Intel Total Memory Encryption and <a href="https://newsroom.intel.com/articles/crypto-acceleration-enabling-path-future-computing">Intel Crypto Acceleration</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/818/1*jRdF16w6ER0oYEgb8vdD4A.png" /></figure><p>Intel SGX is the most researched, updated and battle-tested TEE for data center confidential computing, with the smallest attack surface within the system.</p><p>Intel SGX enables application isolation in private memory regions, called enclaves, to help protect code and data while in use. Intel SGX on our latest 3rd Gen Xeon Scalable processors supports memory enclaves of up to 1 terabyte, designed to help protect even larger datasets and workloads, opening up new possibilities for business transformation in areas such as multiparty collaboration, while helping to maintain data privacy and regulatory compliance.</p><p><strong>Intel Total Memory Encryption</strong></p><p>With growth in data both in the cloud and at the edge, the race is on to encrypt and protect everything from your grocery list to your medical records. To better protect the entire memory of a platform, our 3rd Gen Intel Xeon Scalable Processors introduce Intel Total Memory Encryption (Intel TME). Intel TME helps ensure that all memory accessed from the Intel CPU is encrypted, in order to provide greater protection against hardware physical attacks on the system memory.</p><p>Encryption relies on cryptographic operations that are compute intensive, slowing down user response times and overall system performance. Yet they are necessary because they often support critical business operations where security is paramount, which is why for more than a decade, Intel has led the industry in reducing the compute cost of cryptographic algorithms through innovative instructions, like Intel AES New Instructions (Intel AES-NI), released in 2010.</p><p>The latest 3rd Gen Intel Xeon Scalable processors build upon our leadership position, and Intel Crypto Acceleration delivers breakthrough performance across a host of important cryptographic algorithms, including added acceleration for Secure Hashing Algorithms (SHA), Vectorized AES (VAES), and more. In public-key cryptography we are delivering a 5.6x increase in performance gen over gen along with a 3.3x increase in symmetric encryption².</p><p>Protecting data is essential to extracting value from it, and the capabilities in the 3rd Gen Intel Xeon Scalable platform will help our customers solve their toughest data challenges while improving data confidentiality and integrity. These capabilities are built on top of platforms that are researched, architected, and supported following comprehensive security practices that Intel implements from product concept to end of support. Security doesn’t just happen. It’s the result of unwavering focus resulting in products customers can trust.</p><p>To learn more about how crypto acceleration is helping Tencent accelerate content delivery across their networks to improve their total cost of ownership and how Intel SGX is enabling the Perelman School of Medicine at the University of Pennsylvania, to increase the accuracy of tumor boundaries, check out this video:</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2Fe0TGprBum0M%3Fstart%3D1991&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3De0TGprBum0M&amp;image=http%3A%2F%2Fi.ytimg.com%2Fvi%2Fe0TGprBum0M%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/5e74f9e64a46776d79ffa5e3d7a031c1/href">https://medium.com/media/5e74f9e64a46776d79ffa5e3d7a031c1/href</a></iframe><p><strong>About the Author</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/339/1*kVSynHJZQSQBjPlhFfyFjg.png" /></figure><p>Lisa A. Spelman is corporate vice president in the Data Platforms Group and general manager of the Xeon and Memory Group at Intel Corporation.</p><p>Lisa leads the end-to-end product planning, management, support, and roadmap ownership of Intel® Xeon® processors. Her organization’s responsibilities span strategy and technology, including Intel® Optane™ Persistent Memory enablement and compatibility, across our Xeon platforms.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/807/1*JG-izC4va2nLpO4zMHDEgg.png" /><figcaption>A wafer contains new 3rd Gen Intel Xeon Scalable processors. Intel introduced the processors and the platform they help power on April 6, 2021. Combined with Intel Optane persistent memory and storage, Ethernet adapters, FPGAs and optimized software solutions, the processors deliver performance and workload optimizations across hybrid-cloud, high performance computing, networking and intelligent edge applications. (Credit: Tim Herman/Intel Corporation)</figcaption></figure><p><strong><em>Notices &amp; Disclaimers</em></strong></p><p>¹ Source: <a href="https://www.itpro.com/security/357578/exposed-records-top-36-billion-in-2020-so-far">36 billion personal records exposed by hacks in 2020 so far, by Keumars Afifi-Sabet -29 Oct 2020</a></p><p>² Source: See [70] and [71] at <a href="http://www.intel.com/3gen-xeon-config">www.intel.com/3gen-xeon-config</a>. Results may vary.</p><p>Performance varies by use, configuration and other factors. Learn more at <a href="http://www.intel.com/PerformanceIndex">www.Intel.com/PerformanceIndex</a>​.</p><p>Performance results are based on testing as of dates shown in configurations and may not reflect all publicly available ​updates. See sources cited for configuration details. No product or component can be absolutely secure.</p><p>© 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.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=df326f5e9800" width="1" height="1" alt=""><hr><p><a href="https://medium.com/intel-tech/ice-lake-advances-confidential-computing-with-intel-sgx-total-memory-encryption-df326f5e9800">Ice Lake Advances Confidential Computing with Intel® SGX, Total Memory Encryption</a> was originally published in <a href="https://medium.com/intel-tech">Intel Tech</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AI-powered Automation is Becoming an Increasingly Critical Factor in Industrial Operations]]></title>
            <link>https://medium.com/intel-tech/ai-powered-automation-is-becoming-an-increasingly-critical-factor-in-industrial-operations-d5086bc1c2cc?source=rss-a5e564ccb3ea------2</link>
            <guid isPermaLink="false">https://medium.com/p/d5086bc1c2cc</guid>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[technical-analysis]]></category>
            <category><![CDATA[iot]]></category>
            <category><![CDATA[technews]]></category>
            <category><![CDATA[tech]]></category>
            <dc:creator><![CDATA[Intel Tech]]></dc:creator>
            <pubDate>Wed, 14 Apr 2021 21:19:05 GMT</pubDate>
            <atom:updated>2021-05-11T19:43:09.348Z</atom:updated>
            <content:encoded><![CDATA[<h4><strong>Artificial Intelligence IoT</strong></h4><h4>By Christine Boles, Vice President Internet of Things Group, General Manager, Industrial Solutions Division at Intel Corporation</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*eCIvbXABJQL7RD6T6EiEgQ.jpeg" /></figure><p>The 4th Industrial Revolution (often referred to as Industry 4.0) is ramping up, and many companies are taking full advantage. New technologies introduced by Intel over the last few years give manufacturing companies levels of production insight that were never available before, and provide the opportunity to use ‘smart’ tech to automate their everyday processes.</p><p>Intel has been working directly with ecosystem development partners on innovations that can be incorporated into their specific, real-world solutions and applications. Ultimately, we want everyone to know the full range of Intel’s Market Ready Solutions that can be deployed today.</p><p>For industrial manufacturers, the primary innovations that have had the most positive effect on their operations are advances in artificial intelligence (AI) and machine learning, along with hardware-based visual technology. Using an industrial PC, developers can include the unique capabilities of AI as part of a hybrid machine learning / machine vision application, and then bring their hardware and software applications to integrators who can incorporate them into the appropriate factory lines and processes. Intel is working hard to help manufacturing companies understand the benefits of Intel technologies, what they can expect to be able to do with them, and how they will help to improve their bottom line.</p><p>For many of the companies that Intel has worked with, the primary impetus for the change has been the ongoing pandemic and the business realities that it has forced executives and managers to deal with. At a time when companies are being forced to produce more goods with fewer resources, advancing technology has become an increasingly necessary way to fill the gaps. The concept of being able to remotely manage and have visibility over factory operations wasn’t something that was as critical when everyone was in the factory. Shelter-in-place orders and social distancing requirements, however, have forced many factories to revisit their operations and reevaluate their capabilities. The concept of having that intelligence — that built-in AI learning, machine vision, etc. — in factory operations is now a must have. Companies need the ability to have more of those kinds of IT capabilities built throughout factory operations — even in places they might not have thought they needed as little as a year ago.</p><p>For companies that have not yet made an investment in these new technologies, there is no better time than now to begin. At this point in time, we’ve moved past the debate of “Is this something I should use? Will it bring me benefit?”, along with similar high-level evaluation questions. Factory operators realize that there are benefits to implementing these technologies, whether it’s in monitoring how efficiently their equipment is operating, or collecting data to make process improvements, or even making real-time process decisions as the production line is moving. Enough of the analysis has been done to show that there are significant benefits and return on investment available right now. The next step for those companies is, “How do I deploy this technology, at scale, and introduce those kinds of solutions to my process?” and “How can I build these technologies into the machines themselves?”</p><p>Using industrial PCs equipped with the latest Intel processors, Intel’s OpenVINO toolkit and Edge Insights and Edge Control for Industrial reference designs available on our Edge Software Hub, Intel helps developers, integrators and manufacturers to build customized solutions to ingest time-series data, video, etc.. for particular use cases.</p><p>Often times, when a platform is deployed for a singular purpose, other applications of similar technology in other areas make themselves evident. For more information on this phenomenon, you can read <a href="https://blogs.intel.com/technology/2021/04/edge-learning-is-the-key-to-increasing-automation-capabilities-in-manufacturing/">a piece that I put together for blogs@intel</a>.</p><p><strong><em>Notices &amp; Disclaimers</em></strong></p><p>Performance varies by use, configuration and other factors. Learn more at <a href="http://www.intel.com/PerformanceIndex">www.Intel.com/PerformanceIndex</a>​.</p><p>Your costs and results may vary. Intel technologies may require enabled hardware, software or service activation.</p><p>© 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.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d5086bc1c2cc" width="1" height="1" alt=""><hr><p><a href="https://medium.com/intel-tech/ai-powered-automation-is-becoming-an-increasingly-critical-factor-in-industrial-operations-d5086bc1c2cc">AI-powered Automation is Becoming an Increasingly Critical Factor in Industrial Operations</a> was originally published in <a href="https://medium.com/intel-tech">Intel Tech</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Bringing Edge Insights to IIoT Solutions]]></title>
            <link>https://medium.com/intel-tech/bringing-edge-insights-to-iiot-solutions-5f116ea00cbe?source=rss-a5e564ccb3ea------2</link>
            <guid isPermaLink="false">https://medium.com/p/5f116ea00cbe</guid>
            <category><![CDATA[technology-news]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[internet-of-things]]></category>
            <category><![CDATA[technical-analysis]]></category>
            <category><![CDATA[inteligencia-artificial]]></category>
            <dc:creator><![CDATA[Intel Tech]]></dc:creator>
            <pubDate>Tue, 13 Apr 2021 15:18:01 GMT</pubDate>
            <atom:updated>2021-04-13T15:28:05.561Z</atom:updated>
            <content:encoded><![CDATA[<p><strong>Author: Brian McCarson, Vice President and Senior Principal Engineer at Intel Corporation</strong></p><p>When I started working in Intel’s semiconductor fabs 20 years ago, I was part of the team building some of the early AI capabilities for Intel’s factory process controls systems. Other than large enterprises, mostly in the semiconductor and automotive assembly sectors of the world, few companies could afford to invest in AI because it was so technologically complex and cost prohibitive.</p><p>What is exciting about the time that we are in now is that the community of AI developers is expanding across all regions. We are witnessing the emergence of the AI era and it is shaping up to be one of the fastest adoptions of technological innovations in the history of humankind.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/0*IZ92E96GWuxDMNqh.jpg" /></figure><p>Take for example textile factories, which can now deploy advanced AI vision-based algorithms to perform defect detection on looms that are running at phenomenal speeds in full harmony with the time-series machine data coming off the automated loom machines themselves. Today, equipment that can be purchased and deployed for a few thousand dollars per machine that can improve loom performance, reduce the volume of defective material, and improve factory profit margins. These capabilities that were unheard of even just a few years ago but are now scalable and cost effective.</p><p>Vision-based AI use cases have been the fastest growing area of Intel’s Internet of Things business for several years. All businesses that we support are either investigating or actively deploying AI capabilities. Whether it is automated imaging and CT scan analysis in healthcare, defect detection in pharmaceuticals, predictive maintenance decisions being made off machine data in factories, autonomous driving, or the digital classroom, AI is part of nearly all Internet of Things solutions.</p><p>As I discussed in a previous blog, <a href="https://blogs.intel.com/technology/2021/04/intel-accelerates-ai-for-industrial-applications/">Intel Accelerates AI for Industrial Applications </a>, this is because AI has been simplified, with Python-based or similar low-code tools now accessible for any developer in any domain. Instead of AI exclusively requiring data scientists working on complex, mathematically advanced solutions, software developers have been able to retool and re-skill with open-source software to accelerate business transformations.</p><p>That said, the need for subject matter experts (SMEs) is often overlooked. A data scientist is not going to be able to tell the difference between good textile fabric versus a bad textile fabric. It will be someone with expertise in textiles and loom machinery. As tools become more accessible to the average developer, an engineer with textiles experience who understands Python can find open source algorithms or develop their own and create systems to identify defects in their factory.</p><p>The domain experts in retail, pharmaceuticals, textiles, automotive, and others now can easily harness those tools and apply them to a good business outcome. What the AI community, including Intel, has been doing is delivering to these domain experts access to AI solutions.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*IB4fgfcqRvjw0rHwNO_XVQ.png" /></figure><p>The applications we have been talking about are different than a consumer downloading an app on a smartphone. The application that a large auto company would want to have on their factory floor would preserve their own unique DNA: the unique IP (intellectual property) that they have in their assembly process, and in their design process. They may need help optimizing their applications, but the needs of these end users are much more proprietary.</p><p>They are looking for frameworks so they can deploy applications unique to their business, so that when new IP is introduced , when new products are being introduced to the factory, new government safety protocols are implemented, they can adapt without having to rip up and replace. The same applies when expanding to a new line or even adding a whole new factory. Scaling is also enabled through the same frameworks.</p><p>That’s the vision that these companies are presenting to us, and where Intel is trying to help overcome barriers with our microservices-based enabling tools, such as Intel’s OpenVINO Toolkit combined with Intel’s Edge Insights Software, which uses some of the best open source ingredients in a scalable way to help make today’s applications deployable with an infrastructure that is future-ready. Containerizing applications so they can be moved efficiently from computing resource to computing resource, from the edge to the cloud and back again. Containers are the vehicles that allow customers to move an application from the cloud to an internal network or to an end point device, so their configurations are ready for tomorrow’s challenges.</p><ul><li><a href="https://www.intel.com/content/www/us/en/internet-of-things/industrial-iot/resources/edge-insights-industrial-product-brief.html">Intel’s Edge Insights for Industrial — Product Brief</a></li></ul><p>Intel teams around the world are working with the system integrator community, which plays a big role in the scaling of Industrial IoT solutions. With the democratization of both AI development and AI management, a small SIS (software and information systems) firm with, say, 20 employees, can employ enough data science expertise that they can service dozens of factories in their region with AI services. In fact, the system integrator and the SIS community — which have vertical industry-specific, domain-specific expertise — are getting a lot of traction.</p><ul><li><a href="https://www.intel.co.uk/content/www/uk/en/internet-of-things/industrial-iot/edge-insights-industrial.html">Intel® Edge Insights for Industrial</a></li></ul><p>Take welding, for example. AI can help a weld robot self-calibrate. Instead of having a human do weld quality inspections, an AI algorithm can monitor welds in real-time. The AI algorithm could employ camera inputs if the weld is on the surface or more advanced time-series techniques for spot welds. In microseconds the AI-enhanced control system can determine that a weld is no longer meeting quality standards, and dynamically adjust the controller to change the weld settings. And meanwhile, there has been notification to the maintenance crew that the weld system may need maintenance.</p><p>It takes someone with welding expertise to be able to develop a good welding algorithm, in partnership with someone that has AI capabilities. It is that welding expertise that is key. If it is a large car company which employs data scientists and welding experts, that company might have the capability to put together a weld inspection system with AI capabilities on their own and scale it worldwide to improve the efficiency of their operations.</p><ul><li><a href="https://www.intel.co.uk/content/www/uk/en/customer-spotlight/stories/audi-automated-factory.html">Audi Industry 4.0 Manufacturing Transformation — Intel®</a></li></ul><p>However, what we are finding is that the majority of the world’s businesses don’t have the funding, or even the desire to do that. They would much rather have someone come in with specialized services and be able to say to a customer, “For a monthly subscription fee per tool, I will deploy the capabilities that will generate enough savings to pay for the hardware and the services, and here is my service level agreement.”</p><p>What Intel is doing with our Edge Insights and OpenVINO toolkit-enabled solutions is helping these customers whether they have the ability to invest in data scientists or not. We are finding the pain points in the markets, enabling cloud-to-edge manageability of application services with over-the-air updates, for example. We close the loop so customers have a way to take the world’s most innovative models or algorithms and deploy them at the application level, so they can speak the same language as the existing infrastructure on the factory floor.</p><p>Intel is exceptionally talented at control systems infrastructure and has been fueling the industrial PC sector for decades. The control system infrastructure is ideally suited for traditional CPU compute cycles, and by combining support for AI capabilities in harmony with Intel’s integrated GPUs and VPU acceleration capabilities, it is starting to shift ideas to reality.</p><p>To learn more about Intel software designs for IIoT watch this tutorial on <a href="https://www.intel.com/content/www/us/en/internet-of-things/industrial-iot/edge-insights-industrial.html">Intel® Edge Insights for Industrial</a>.</p><p><strong><em>About the Author</em></strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/391/1*s27wQYljTQoTN0Ct8TkacA.png" /><figcaption>Brian McCarson VP and Senior Principal Engineer in the Internet of Things Group</figcaption></figure><p>Brian McCarson is a Vice President and Senior Principal Engineer in the Internet of Things Group. He manages the Industrial Solutions Division Systems Engineering &amp; Architecture Team and specializes in deployment of AI and controls systems for Industrial applications.</p><p><strong><em>Notices &amp; Disclaimers</em></strong></p><p>Performance varies by use, configuration and other factors. Learn more at <a href="http://www.intel.com/PerformanceIndex">www.Intel.com/PerformanceIndex</a>​.</p><p>Your costs and results may vary. Intel technologies may require enabled hardware, software or service activation.</p><p>© 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.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5f116ea00cbe" width="1" height="1" alt=""><hr><p><a href="https://medium.com/intel-tech/bringing-edge-insights-to-iiot-solutions-5f116ea00cbe">Bringing Edge Insights to IIoT Solutions</a> was originally published in <a href="https://medium.com/intel-tech">Intel Tech</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The evolution of XPU and the critical role of software]]></title>
            <link>https://medium.com/intel-tech/the-evolution-of-xpu-and-the-critical-role-of-software-c46970dfcbe9?source=rss-a5e564ccb3ea------2</link>
            <guid isPermaLink="false">https://medium.com/p/c46970dfcbe9</guid>
            <category><![CDATA[tech]]></category>
            <category><![CDATA[technews]]></category>
            <category><![CDATA[nvidia]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[technical-analysis]]></category>
            <dc:creator><![CDATA[Intel Tech]]></dc:creator>
            <pubDate>Mon, 12 Apr 2021 15:17:34 GMT</pubDate>
            <atom:updated>2022-07-14T17:30:42.943Z</atom:updated>
            <content:encoded><![CDATA[<h4>Contributors: Jason Lawley, Director Technology Leadership Marketing and Chandan Damannagari, Software Technology Expert at Intel</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/623/1*vxDor7VN0T20DNRdM51D1g.png" /><figcaption>Figure 1.0 Taxonomy of modern workloads</figcaption></figure><p>The long-predicted data explosion is here, and one thing is clear — the days of a single architecture to process data are behind us. The scale of the explosion can be seen in the ~1GB/day of data generated by an individual internet user, 50GB/day by a smart car, 3TB/day by a smart hospital, and 50PB/day by a smart city¹. Not only is the amount and diversity of data today staggering, but also where, when, and how it’s processed is rapidly evolving. Whether a workload is run on the edge or in the cloud, or if it is an AI workload or a storage workload, having the right architecture and software to utilize that architecture can make or break a project. At Intel, we are on a journey to provide every developer with the right combination of hardware architectures and software to build the optimal solution for their particular use case. We are doing this both through our XPU strategy and by enabling an open, standards-based cross-architecture programming model called oneAPI. But what is an XPU? It’s not a new processor or product, but a portfolio of architectures (CPU, GPU, FPGA and other accelerators). This blog explains the concept of XPU and how software is critical to the success of Intel’s XPU future.</p><p><strong>THE NEED FOR DIVERSE COMPUTING ARCHITECTURES</strong></p><p>If raw data is the proverbial ore, then data processing is how gold is extracted from that raw data. Processing the most amount of data with the lowest power has always been the goal of most compute architectures. In recent years, an important aspect to this gold analogy is <em>time</em>. The value of data is fleeting and often without processing that data at the right time, it becomes useless. Think of things like real-time ad services, automotive sensing, or live video transcoding. They are all sensitive to how long it takes to process the data. With this in mind, there are a few key characteristics that developers look for in their hardware and that computer architects strive to deliver. They are high performance, low latency, and low power.</p><ul><li>Performance — The amount of data that can be processed per unit of time. This might be video analytics or encoding, searching a database, or a variety of other data processing activities. Performance might be measured in GByte/s or images/s.</li><li>Latency — The length of time required to process the data and to get a result. In some applications, latency is not an issue while in others like real-time digital assistants or automotive video processing, it is a critical factor. Latency is often measured in milli-seconds or even micro-seconds.</li><li>Power — The amount of power needed to process the data. Power is important both at the edge and in the cloud. As devices move to the edge, where only batteries or other limited power sources are available, power consumption rates are increasingly critical. Even in the data center, cooling and electricity costs constitute a significant percentage of total operating costs. Power is often measured in watts or joules.</li></ul><p>Unfortunately, there is no longer a one size fits all architecture for the diversity of today’s data. Specialized architectures have shown that they can optimize performance, power, or latency as needed, but getting the right balance for a gaming GPU is vastly different than what is needed for a battery-operated security camera. Architectures also continue to evolve as the types of data and workloads diversify. The ability to create a broad set of architectures targeting an even broader set of workloads is the essence of Intel’s XPU strategy.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FdXDP-SJ0e2o&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DdXDP-SJ0e2o&amp;image=http%3A%2F%2Fi.ytimg.com%2Fvi%2FdXDP-SJ0e2o%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/8a243fc9ece20c963a53de09effd2a43/href">https://medium.com/media/8a243fc9ece20c963a53de09effd2a43/href</a></iframe><p><strong>XPU AND oneAPI — TWO PIECES OF THE SAME HETEROGENEOUS COMPUTING PUZZLE</strong></p><p>Having the right architecture for the right application is critical to our customers success. CPUs and the vast applicability of their general-purpose computing capabilities are central to our <a href="https://cqpreview.intel.com/content/www/us/en/architecture-and-technology/xpu.html">XPU strategy</a>. We also offer an already successful lineup of GPUs, FPGAs, VPUs, and other architecture types to meet the processing needs of the huge variety of different workloads that exist today and into the future. We classified the range of architectures into Scalar, Vector, Matrix and Spatial as we explained in <a href="https://blogs.intel.com/technology/2020/08/harnessing-the-power-of-a-heterogeneous-computing-future/#gs.ye982a">Harnessing the Power of a Heterogenous Computing Future.</a> Our mission is to offer the broadest suite of XPU silicon in the industry, to enable our customers and communities in this next era of computing.</p><p>Having multiple architectures is just one piece of the puzzle though. The bigger challenge is to quickly determine which architecture makes the most sense for a desired workload. Often, the best way to evaluate a workload is to try that workload on the different architectures and compare. The migration process between architectures can be time consuming and error prone. Enter oneAPI, a cross-industry, open, standards-based unified programming model. Using oneAPI, developers are given the freedom to choose the best architectures to meet their workload/application needs, while using a single code base across multiple architectures. This allows developers to maximize cross-architecture performance and minimize development costs, while also giving them the freedom to expose and exploit cutting-edge features across diverse XPU architectures. By abstracting hardware heterogeneity with oneAPI, Intel is driving a new era of accelerated computing and helping to best position our customers to harness the power of XPUs now and into the future.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/928/1*lHYdFb6KEvrXYFENb20heg.png" /><figcaption>Figure 2.0 XPU and OneAPI</figcaption></figure><p>This is the <strong>first blog in a series of blogs</strong> where we will explore the scalar, vector, matrix and spatial architectures in greater detail. You can also check out this <a href="https://software.intel.com/content/www/us/en/develop/articles/xpublog0001-hello-sycl-and-dpcpp.html">blog</a> from James Reinders to learn more about how DPC++ and oneAPI can help unleash the power of XPUs. We have made significant strides on our journey to become an XPU company and we continue to strive to provide class-leading solutions to harness the limitless potential of the world’s data.</p><p><strong><em>Notes and Disclaimers</em></strong></p><p>¹ source: <a href="https://www.cisco.com/c/dam/m/en_us/service-provider/ciscoknowledgenetwork/files/547_11_10-15-DocumentsCisco_GCI_Deck_2014-2019_for_CKN__10NOV2015_.pdf">https://www.cisco.com/c/dam/m/en_us/service-provider/ciscoknowledgenetwork/files/547_11_10-15-DocumentsCisco_GCI_Deck_2014-2019_for_CKN__10NOV2015_.pdf</a></p><p>© Intel Corporation. Intel, the Intel logo, and other Intel marks are trademarks of Intel Corporation or its subs. Performance varies by use, configuration and other factors. Learn more at <a href="http://www.intel.com/PerformanceIndex">www.Intel.com/PerformanceIndex</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c46970dfcbe9" width="1" height="1" alt=""><hr><p><a href="https://medium.com/intel-tech/the-evolution-of-xpu-and-the-critical-role-of-software-c46970dfcbe9">The evolution of XPU and the critical role of software</a> was originally published in <a href="https://medium.com/intel-tech">Intel Tech</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>