Slashdot reader katydid77 shares a report from the supercomputing site The Next Platform: If you need any proof that it doesn't take the most advanced chip manufacturing processes to create an exascale-class supercomputer, you need look no further than the Sunway "OceanLight" system housed at the National Supercomputing Center in Wuxi, China. Some of the architectural details of the OceanLight supercomputer came to our attention as part of a paper published by Alibaba Group, Tsinghua University, DAMO Academy, Zhejiang Lab, and Beijing Academy of Artificial Intelligence, which is running a pretrained machine learning model called BaGuaLu, across more than 37 million cores and 14.5 trillion parameters (presumably with FP32 single precision), and has the capability to scale to 174 trillion parameters (and approaching what is called "brain-scale" where the number of parameters starts approaching the number of synapses in the human brain)....

Add it all up, and the 105 cabinet system tested on the BaGuaLu training model, with its 107,250 SW26010-Pro processors, had a peak theoretical performance of 1.51 exaflops. We like base 2 numbers and think that the OceanLight system probably scales to 160 cabinets, which would be 163,840 nodes and just under 2.3 exaflops of peak FP64 and FP32 performance. If it is only 120 cabinets (also a base 2 number), OceanLight will come in at 1.72 exaflops peak. But these rack scales are, once again, just hunches. If the 160 cabinet scale is the maximum for OceanLight, then China could best the performance of the 1.5 exaflops "Frontier" supercomputer being tuned up at Oak Ridge National Laboratories today and also extend beyond the peak theoretical performance of the 2 exaflops "Aurora" supercomputer coming to Argonne National Laboratory later this year — and maybe even further than the "El Capitan" supercomputer going into Lawrence Livermore National Laboratory in 2023 and expected to be around 2.2 exaflops to 2.3 exaflops according to the scuttlebutt.

We would love to see the thermals and costs of OceanLight. The SW26010-Pro chip could burn very hot, to be sure, and run up the electric bill for power and cooling, but if SMIC [China's largest foundry] can get good yield on 14 nanometer processes, the chip could be a lot less expensive to make than, say, a massive GPU accelerator from Nvidia, AMD, or Intel. (It's hard to say.) Regardless, having indigenous parts matters more than power efficiency for China right now, and into its future, and we said as much last summer when contemplating China's long road to IT independence. Imagine what China can do with a shrink to 7 nanometer processes when SMIC delivers them — apparently not even using extreme ultraviolet (EUV) light — many years hence....

The bottom line is that the National Research Center of Parallel Computer Engineering and Technology (known as NRCPC), working with SMIC, has had an exascale machine in the field for a year already. (There are two, in fact.) Can the United States say that right now? No it can't.

Apple has finally discontinued the 27-inch iMac. Minutes after Apple's Peek Performance event wrapped, the last of the Intel iMacs disappeared from its online store.

In a 15-minute-long video, YouTuber Quinn Nelson from Snazzy Labs explains how he managed to shrink the current M1 Mac Mini by 78% without harming performance. 9to5Mac reports: In conclusion, by rearranging the internals and swapping out the power supply, Nelson was able to reduce the size of the Mac mini enclosure by 78%. He organized all the parts inside a 3D-printed body with a mini Mac Pro motif.

The reason that theoretical space savings are so huge is because when Apple released the first round of Apple Silicon computers, they did not change the hardware industrial design at all. So the current Mac Mini enclosure is designed to fit an Intel CPU and circuit board, including having to accommodate the large fans and heat sinks the Intel chip required.

But with the power efficiency of the M1, Apple has the headroom to do something much more drastic. Indeed, a lot of the M1 Mac mini internals is just empty space. The Snazzy Labs video gives a glimpse at what is possible if Apple is more ambitious with the next-generation Mac mini design, and tries to create something truly mini. The CAD files and schematics can be viewed here.

A number of industry stalwarts including Intel, AMD, Arm, TSMC, and Samsung on Wednesday introduced a new Universal Chiplet Interconnect Express (UCIe) consortium. AnandTech: Taking significant inspiration from the very successful PCI-Express playbook, with UCIe the involved firms are creating a standard for connecting chiplets, with the goal of having a single set of standards that not only simplify the process for all involved, but lead the way towards full interoperability between chiplets from different manufacturers, allowing chips to mix-and-match chiplets as chip makers see fit. In other words, to make a complete and compatible ecosystem out of chiplets, much like today's ecosystem for PCIe-based expansion cards.

The comparisons to PCIe are apt on multiple levels, and this is perhaps the best way to quickly understand the UCIe group's goals. Not only is the new standard being made available in an open fashion, but the companies involved will be establishing a formal consortium group later this year to administer UCIe and further develop it. Meanwhile from a general technology perspective, the use of chiplets is the latest step in the continual consolidation of integrated circuits, as smaller and smaller transistors have allowed more and more functionality to be brought on-chip. In essence, features that have been on an expansion card or separate chip up until now are starting to make their way on to the chip/SoC itself. So like PCIe moderates how these parts work together as expansion cards, a new standard has become needed to moderate how these parts should work together as chiplets.

President Joe Biden called on Congress to pass the CHIPS Act, a law that would provide chipmakers with $52 billion in subsidies to advance semiconductor manufacturing in the United States, during his State of the Union speech Tuesday. From a report: Biden lauded Intel Chief Executive Pat Gelsinger, who last month announced a $20 billion investment for two new chip fabrication facilities, or fabs, that the company will build just west of Columbus, Ohio. Intel plans to spend $100 billion to build the Ohio "megafab" over the next decade, with an eventual total of eight fabs, but the speed of that investment will depend on the US subsidy, Gelsinger has said.

"Intel's CEO, Pat Gelsinger, who is here tonight, told me they are ready to increase their investment from $20 billion to $100 billion. That would be one of the biggest investments in manufacturing in American history," Biden said. "And all they're waiting for is for you to pass this bill. ... Send it to my desk. I'll sign it." The Senate passed a bill funding the CHIPS Act in 2021, and the House of Representatives followed suit in February, but the differences in the bills haven't been ironed out in committee and the subsidy hasn't arrived despite some bipartisan support. The funding would help the US compete with government help in Taiwan and South Korea, where leading chipmakers Taiwan Semiconductor Manufacturing Co. (TSMC) and Samsung have the bulk of their operations. The US subsidies would knock about $3 billion off the $10 billion price tag for a new fab, a subsidy level Intel says matches those in Asia.

An anonymous reader quotes a report from Reuters: Italy plans to set aside more than $4.6 billion until 2030 to boost domestic chip manufacturing as it seeks to attract more investment from tech companies such as Intel, a draft decree seen by Reuters showed on Tuesday. The government is trying to persuade the U.S. group to spend billions of euros on an advanced chipmaking plant in Italy that uses innovative technologies to weave full chips.

Rome is ready to offer Intel public money and other favorable terms to fund part of the overall investment, which is expected to be worth around $9 billion over 10 years, Reuters reported in December. To boost domestic chipmaking, Italy is also in talks with French-Italian STMicroelectronics , Taiwanese-controlled MEMC Electronic Materials Inc and Israeli Tower Semiconductor, which is set to be bought by Intel. Negotiations with Intel are complex as the U.S. group has tabled very tough demands, a government source involved in the talks told Reuters.

As part of an 8 billion euro package to support the economy and curb surging energy bills, Italy plans to allocate 150 million euros in 2022 and 500 million euros per year from 2023 until 2030, the decree showed. The Italian government will promote "research and development of microprocessor technology and investments in new industrial applications of innovative technologies," the legislation added. Rome aims to use the funding also to convert existing industrial sites and favor the construction of new plants in Italy.

An anonymous reader shares a report: Lenovo has dumped a whole bunch of new ThinkPads into the world, and there's some exciting stuff in there. We're getting a brand-new ThinkPad X13s powered by Snapdragon chips, a fifth-generation ThinkPad X1 Extreme with a WQXGA 165Hz screen option, and new additions to the P-series and T-series as well. The news I'm personally most excited about is the screen shape. A few months ago, Lenovo told me that much of its portfolio would be moving to the 16:10 aspect ratio this year. They appear to be keeping their word. Across the board, the new models are 16:10 -- taller and roomier than they were in their 16:9 eras. Some news that's a bit more... intriguing is the all-new ThinkPad X13s, which is the first laptop to feature the Snapdragon 8cx Gen 3 compute platform. Qualcomm made some lofty claims about this platform upon its release, including "60 percent greater performance per watt" over competing x86 platforms and "multi-day battery life." The ThinkPad X13s will run an Arm version of Windows 11, with its x64 app emulation support. The P-series models and Intel T-series models will all be here in April, with prices ranging from $1,399 to $1,419.

Russia's invasion of Ukraine by land, air and sea risks reverberating across the global chip industry and exacerbating current supply-chain constraints. Reuters Breakingviews: Ukraine is a major producer of neon gas critical for lasers used in chipmaking and supplies more than 90% of U.S. semiconductor-grade neon, according to estimates from research firm Techcet. About 35% of palladium, a rare metal also used for semiconductors, is sourced from Russia. A full-scale conflict disrupting exports of these elements might hit players like Intel, which gets about 50% of its neon from Eastern Europe, according to JPMorgan. ASML, which supplies machines to semiconductor makers, sources less than 20% of the gases it uses from the crisis-hit countries.

Hmmmmmm shares a report from Yahoo Finance: AMD's market cap currently stands at $188 billion after shares rose nearly 2% in Tuesday's session. Intel's market cap is $182 billion. That marks the second time in a week AMD's market value has climbed above Intel -- the first time it happened was a week ago. Followers of this battle may not be surprised to see this one happen (and seeing it continue from here) for several reasons. First, AMD has been winning the battle on Wall Street for sexier investment thesis. AMD last week closed on its $35 billion acquisition for Xilinx. Secondarily, AMD has flat out posted better financials than Intel (for some time) as it has gained market share in key areas (notably in servers). AMD's sales and profits rose 68% and 117%, respectively in 2021. The company outlined 31% revenue growth for 2022 and gross profit margins of 51%. Intel's 2021 sales and earnings increased 2% and 7%, respectively. The company sees sales in 2022 rising about 2%. Profits are expected to drop 36% as Intel further builds out its chip-making capacity.

Intel has acquired Linutronix, the German-based Linux consulting firm that is focused on embedded Linux and real-time computing. From a report: Intel's acquisition of Linutronix appears to be primarily focused as an acqui-hire with getting Linutronix's very talented staff at Intel. Among the prominent Linutronix engineers is their CTO Thomas Gleixner as a longtime kernel maintainer and important contributor on the x86 side, including with Linux's CPU security mitigations and perhaps most notably for the real-time (PREEMPT_RT) work.

Intel launched the first wave of its 12th Gen Alder Lake chips at CES 2022 -- but only for its H-series lineup of chips, destined for the most powerful and power-hungry laptops. And now, it's rolling out the rest of its Alder Lake laptop lineup: the P-series and U-series models it briefly showed off in January, which are set to power the thinner, lighter, and cheaper laptops of 2022. From a report: In total, there are a whopping 20 chips fit for a wide range of hardware across the P-series, U-series (15W), and U-series (9W) categories, with the first laptops powered by the new processors set to arrive in March. Like their more powerful H-series cousins (and the Alder Lake desktop chips that Intel launched in late 2021 and at CES 2022), the new P-series and U-series chips have a lot more cores than 2020's 11th Gen models, with a hybrid architecture approach that combines performance and efficiency cores to maximize both power and battery life. And Intel is promising some big improvements focused around those boosted core counts, touting up to 70 percent better multi-thread performance than previous 11th Gen (and AMD) hardware. The company also says that it wins out in benchmarks against chips like Apple's M1 and M1 Pro (although not the M1 Max), and AMD's Ryzen R7 5800U in tasks like web browsing and photo editing.

Protocol reports that Qualcomm will finally jump on the AV1 video codec bandwagon next year. Ars Technica reports: AV1 is the web's next open, royalty-free video codec, and widespread adoption will require hardware support from the world's chip vendors. Qualcomm's 2022 flagship SoC, the Snapdragon 8 Gen 1 chip, doesn't support AV1. Samsung's Exynos 2200 managed to ship the video codec this year in international versions of the Galaxy S22, while the MediaTek Dimensity 1000 SoC has been shipping in phones for over a year now with AV1 support. Apple is a founding member of the AV1 Alliance, but its devices also don't support the codec yet.

The report says Qualcomm's "upcoming flagship Snapdragon mobile processor" -- model number "SM8550" -- will support AV1. That would probably be called the "Snapdragon 8 Gen 2" SoC, due out in 2023. Wide adoption of AV1 seems inevitable, though it is taking a while. The codec is a successor to Google's VP8 and VP9 codecs and is being built by the Alliance for Open Media. The alliance's lineup is a who's who of tech companies, with founding members like Amazon, Apple, ARM, Facebook, Google, Intel, Microsoft, Mozilla, Netflix, Nvidia, and Samsung. Netflix and Google's YouTube are both making AV1 support "a requirement" for future products that want to support either video service. That should motivate just about every hardware and software vendor out there to get the job done. Aside from being open source and royalty-free, the report notes that the newer AV1 codec also has the benefit of being 30% more efficient than H.265.

AnandTech reports: It's no secret that Intel's enterprise processor platform has been stretched in recent generations. Compared to the competition, Intel is chasing its multi-die strategy while relying on a manufacturing platform that hasn't offered the best in the market. That being said, Intel is quoting more shipments of its latest Xeon products in December than AMD shipped in all of 2021, and the company is launching the next generation Sapphire Rapids Xeon Scalable platform later in 2022. Beyond Sapphire Rapids has been somewhat under the hood, with minor leaks here and there, but today Intel is lifting the lid on that roadmap.

Currently in the market is Intel's Ice Lake 3rd Generation Xeon Scalable platform, built on Intel's 10nm process node with up to 40 Sunny Cove cores. The die is large, around 660 mm2, and in our benchmarks we saw a sizeable generational uplift in performance compared to the 2nd Generation Xeon offering. The response to Ice Lake Xeon has been mixed, given the competition in the market, but Intel has forged ahead by leveraging a more complete platform coupled with FPGAs, memory, storage, networking, and its unique accelerator offerings. Datacenter revenues, depending on the quarter you look at, are either up or down based on how customers are digesting their current processor inventories (as stated by CEO Pat Gelsinger). Further reading: Intel Arc Update: Alchemist Laptops Q1, Desktops Q2; 4M GPUs Total for 2022.

As part of Intel's annual investor meeting taking place today, Raja Koduri, Intel's SVP and GM of the Accelerated Computing Systems and Graphics (AXG) Group delivered an update to investors on the state of Intel's GPU and accelerator group, including some fresh news on the state of Intel's first generation of Arc graphics products. AnandTech: Among other things, the GPU frontman confirmed that while Intel will indeed ship the first Arc mobile products in the current quarter, desktop products will not come until Q2. Meanwhile, in the first disclosure of chip volumes, Intel is now projecting that they'll ship 4mil+ Arc GPUs this year. In terms of timing, today's disclosure confirms some earlier suspicions that developed following Intel's CES 2022 presentation: that the company would get its mobile Arc products out before their desktop products. Desktop products will now follow in the second quarter of this year, a couple of months behind the mobile parts. And finally, workstation products, which Intel has previously hinted at, are on their way and will land in Q3.

As Masayoshi Son tried to persuade investors of the wisdom of purchasing one of the most successful chip companies in the world in 2016, the SoftBank chief had one clear message: "For the era of the 'Internet of things,' I think the champion will be Arm." But the concept of connecting billions of everyday and industrial devices to the Internet has been much slower than anticipated to materialize. From a report: Son's drive to capture the chip design market for the Internet of things (IoT) was the first bet he made on Arm that has not paid off. The second was a $66 billion sale of the company to Nvidia that unraveled last week. Arm remains the dominant player in designing chips for smartphones, still the most ubiquitous form of computing but a source of much slower growth in recent years. Ahead of an initial public offering that could come as soon as this year, the company is racing to solidify its position in new markets that it has underexploited to date, while trying to drive up profits to appeal to a new set of investors. Rene Haas, Arm's incoming chief executive, told the Financial Times that its products were now "far more competitive" in data centers and cars than when SoftBank bought the Cambridge-based company.

"Making trade-offs about where to invest, where not to invest...âthose are the trade-offs that public companies and even private companies have to do every day," he said. "The company is in great shape." When Son spearheaded the $31 billion purchase of Arm, he saw it as a wager on the future of the entire technology industry, which was crystallizing at that time around the IoT concept. He proceeded to push the executive team firmly on the course to designing chips for this future of machine connectivity. Five-and-a-half years later, it has become increasingly clear that the IoT gamble was a costly misadventure. Moreover, it distracted Arm from attacking Intel's dominance in the much larger data center market. As Son's vision collided with reality, SoftBank quietly revised its market calculations. A presentation from 2018 forecast that by 2026, the IoT controller market would be worth $24 billion, and the server market $22 billion. But, a similar presentation from 2020 predicted that by 2029, the IoT chip market would reach only $16 billion, while the server market -- of which Arm had so far only captured a 5 percent share -- would reach $32 billion. The Japanese technology group also revised down its estimate of the value of the IoT market, from $7 billion in 2017 to $4 billion in 2019.

In an effort to boost its manufacturing operations, Intel is nearing a deal to buy Israeli chip firm Tower Semiconductor for almost $6 billion, reports Reuters, citing a report from the Wall Street Journal. "A deal could be unveiled as soon as this week assuming the talks do not fall apart, the report said, citing people familiar with the matter." It would likely include a hefty premium, given Tower's market value of roughly $3.6 billion. Following the news, shares of Tower Semiconductor, which specializes in making analog integrated circuits, were up nearly 53% in extended trading while Intel was down 0.8%.

The Linux 5.18 kernel is adding support this spring for the Intel Hardware Feedback Interface to make better decisions about where to place given work among available CPU cores/threads, reports Phoronix.

This is significant because Intel's Alder Lake CPUs "are the first x86-64 processors to embrace a hybrid paradigm with two separate CPU architectures on the same die," explains Hot Hardware: These two separate CPU architectures have different strengths and capabilities. The Golden Cove "performance cores" (or P-cores) feature Intel's latest high-performance desktop CPU architecture, and they are blisteringly fast. Meanwhile, the Gracemont "efficiency cores" (or E-cores) are so small that four of them, along with 2MB of shared L2 cache, can nearly fit in the same space as a single Golden Cove core. They're slower than the Golden Cove cores, but also much more efficient, at least in theory.

The idea is that background tasks and light workloads can be run on the E-cores, saving power, while latency-sensitive and compute-intensive tasks can be run on the faster P-cores. The benefits of this may not have been exactly as clear as Intel would have liked on Windows, but they were even less visible on Linux. That's because Linux isn't aware of the unusual configuration of Alder Lake CPUs.

Well, that's changing in Linux 5.18, slated for release this spring. Linux 5.18 is bringing support for the Intel Enhanced Hardware Feedback Interface, or EHFI...

This is essentially the crux of Intel's "Thread Director," which is an intelligent, low-latency hardware-assisted scheduler.

Intel is entering the blockchain mining market with an upcoming GPU capable of mining Bitcoin. From a report: Intel insists the effort won't put a strain energy supplies or deprive consumers of chips. The goal is to create the most energy-efficient blockchain mining equipment on the planet, it says. "We expect that our circuit innovations will deliver a blockchain accelerator that has over 1,000x better performance per watt than mainstream GPUs for SHA-256 based mining," Intel's General Manager for Graphics, Raja Koduri, said in the announcement. (SHA-256 is a reference to the mining algorithm used to create Bitcoins.)

News of Intel's blockchain-mining effort first emerged last month after the ISSCC technology conference posted details about an upcoming Intel presentation titled: "Bonanza Mine: An Ultra-Low-Voltage Energy-Efficient Bitcoin Mining ASIC." ASICs are chips designed for a specific purpose, and also refer to dedicated hardware to mine Bitcoin. Friday's announcement from Koduri added that Intel is establishing a new "Custom Compute Group" to create chip platforms optimized for customers' workloads, including for blockchains.

Intel's mysterious Software Defined Silicon (SDSi) mechanism for adding features to Xeon CPUs will be officially supported in Linux 5.18, the next major release of the operating system. Tom's Hardware reports: SDSi allows users to add features to their CPU after they've already purchased it. Formal SDSi support means that the technology is coming to Intel's Xeon processors that will be released rather shortly, implying Sapphire Rapids will be the first CPUs with SDSi. Intel started to roll out Linux patches to enable its SDSi functionality in the OS last September. By now, several sets of patches have been released and it looks like they will be added to Linux 5.18, which is due this Spring. Hans de Goede, a long-time Linux developer who works at Red Hat on a wide array of hardware enablement related projects, claims that SDSi will land in Linux 5.18 if no problems emerge, reports Phoronix. "Assuming no major issues are found, the plan definitely is to get this in before the 5.18 merge window," said de Goede.

Intel Software Defined Silicon (SDSi) is a mechanism for activating additional silicon features in already produced and deployed server CPUs using the software. While formal support for the functionality is coming to Linux 5.18 and is set to be available this spring, Intel hasn't disclosed what exactly it plans to enable using its pay-as-you-go CPU upgrade model. We don't know how it works and what it enables, but we can make some educated guesses. [...]

An anonymous reader quotes a report from Motherboard: A section of the New York Police Department (NYPD) focused on intelligence gathering received a demo of NSO Group's controversial Pegasus spyware product, according to an email obtained by Motherboard. The news provides more insight into Israeli company NSO Group's push into the surveillance market in the United States, and specifically its pitching of the company's technology to American police forces. The findings come after the New York Times reported that the FBI bought a Pegasus license in 2019 for evaluation purposes.

"There will be a demo of the attached investigative software at the Rutgers School of Criminal Justice," James Sheehan, a program manager from the Northern New Jersey-Newark/Jersey City UASI, wrote in the August 2015 email. The UASI is the Urban Area Security Initiative, a program administered by the Department of Homeland Security which brings together bodies from law enforcement, fire service, public health, and more to address threats of terrorism and other issues. "The audience is the UASI/CorrStat region and NYPD intel," Sheehan continued. Recipients on Sheehan's email inviting people to attend included representatives from the Bergen County Prosecutor's Office, Jersey City's public safety agency, and the Paterson Police Department, a city of just over 150,000.

Attached to Sheehan's email was a brochure for Pegasus, NSO Group's hacking product, which advertised the tool's ability to obtain a target's calls, contacts, emails, WhatsApp messages, track their location, and more. The brochure contains a logo for WestBridge, NSO Group's North American branch. "Turn Your Target's Smartphone into an Intelligence Gold Mine," the Pegasus brochure reads. "NYPD intel" likely refers to the NYPD's Intelligence Bureau. Its mission is to "detect and disrupt criminal and terrorist activity through the use of intelligence-led policing. In combination with traditional policing methods, uniformed officers and civilian analysts in the Intelligence Bureau collect and analyze information from a variety of sources in order to advance criminal and terrorist investigations," according to the NYPD's website.