Nvidia announced yesterday that its upcoming H100 "Hopper" Tensor Core GPU set new performance records during its debut in the industry-standard MLPerf benchmarks, delivering results up to 4.5 times faster than the A100, which is currently Nvidia's fastest production AI chip. Ars Technica reports: The MPerf benchmarks (technically called "MLPerfTM Inference 2.1") measure "inference" workloads, which demonstrate how well a chip can apply a previously trained machine learning model to new data. A group of industry firms known as the MLCommons developed the MLPerf benchmarks in 2018 to deliver a standardized metric for conveying machine learning performance to potential customers. In particular, the H100 did well in the BERT-Large benchmark, which measures natural language-processing performance using the BERT model developed by Google. Nvidia credits this particular result to the Hopper architecture's Transformer Engine, which specifically accelerates training transformer models. This means that the H100 could accelerate future natural language models similar to OpenAI's GPT-3, which can compose written works in many different styles and hold conversational chats. The chip, which is still in development, is forecasted to replace the A100 as the company's flagship data center GPU. The U.S. government imposed restrictions last week on exports of the chips to China, leading many to fear that Nvidia might not be able to deliver the H100 by the end of the year. However, Nvidia clarified that the U.S. government will allow it to continue development of the chip in China, "so the project appears back on track for now," reports Ars.

Garmin has reacted to Apple's new rugged Apple Watch Ultra, saying in a tweet following the iPhone 14 and Apple Watch event that it measures battery life in "months" and "not hours," promoting its latest Enduro 2 watch for athletes. MacRumors reports: While the Apple Watch Ultra has the longest battery life of any Apple Watch to date, with Apple promising up to 36 hours of normal use and up to 60 hours with watchOS 9's new Low Power Mode setting and other optimizations, the Enduro 2 can last significantly longer depending on usage scenarios. Despite Garmin's claim that it measures battery life in months, the company actually advertises the Enduro 2 as having "up to 150 hours of battery life in GPS mode with solar charging" and "up to 34 days of battery life in smartwatch mode." The Enduro 2 has a 1.4-inch solar-powered display, compared to the nearly 2-inch display on the Apple Watch Ultra that can reach a peak brightness of 2,000 nits, the brightest ever in an Apple Watch. Battery life seems to be the main point of contention, based on comments from Hacker News and Reddit. "As someone who is into trail running and backpacking, a 36 hour battery life is untenable, and charging it every night seems like a pain," writes user lukeinator42.

"The real irony of the charge-at-night model is that you miss the single biggest source of error in human life: lack of sleep," adds killjoywashere. "Accurate sleep measurement is, on a day-to-day basis, far more valuable than many of the features advertised on this system."

An anonymous reader quotes a report from The Register: The biz's GENE-ADP6, announced this week, can pack as much as a 12-core/16-thread Intel processor with Iris Xe graphics into a 3.5-inch form factor. The diminutive system is aimed at machine-vision applications and can be configured with your choice of Intel silicon including Celeron, or Core i3, i5, or a choice of 10 or 12-core i7 processors. As with other SBCs we've seen from Aaeon and others, the processors aren't socketed so you won't be upgrading later. This device is pretty much aimed at embedded and industrial use, mind. All five SKUs are powered by Intel's current-gen Alder Lake mobile processor family, including a somewhat unusual 5-core Celeron processor that pairs a single performance core with four efficiency cores. However, only the i5 and i7 SKUs come equipped with Intel's Iris Xe integrated graphics. The i3 and Celeron are stuck on UHD graphics. The board can be equipped with up to 64GB of DDR5 memory operating at up to 4800 megatransfers/sec by way of a pair of SODIMM modules.

For I/O the board features a nice set of connectivity including a pair of NICs operating at 2.5 Gbit/sec and 1 Gbit/sec, HDMI 2.1 and Display Port 1.4, three 10Gbit/sec-capable USB 3.2 Gen 2 ports, and a single USB-C port that supports up to 15W of power delivery and display out. For those looking for additional connectivity for their embedded applications, the system also features a plethora of pin headers for USB 2.0, display out, serial interfaces, and 8-bit GPIO. Storage is provided by your choice of a SATA 3.0 interface or a m.2 mSATA/NVMe SSD. Unlike Aaeon's Epic-TGH7 announced last month, the GENE-ADP6 is too small to accommodate a standard PCIe slot, but does feature a FPC connector, which the company says supports additional NVMe storage or external graphics by way of a 4x PCIe 4.0 interface.

The state of Virginia has over a third of America's hyperscale datacenter capacity, and this amounts to more than the entire capacity of China or the whole of Europe, highlighting just how much infrastructure is concentrated along the so-called Datacenter Alley. The Register reports: These figures come from Synergy Research Group, which said that the US accounts for 53 percent of global hyperscale datacenter capacity, as measured by critical IT load, at the end of the second quarter of 2022. The remainder is relatively evenly split between China, Europe, and the rest of the world. While few would be surprised at the US accounting for the lion's share of datacenter capacity, the fact that so much is concentrated in one state could raise a few eyebrows, especially when it is centered on a small number of counties in Northern Virginia -- typically Loudoun, Prince William, and Fairfax -- which make up Datacenter Alley.

"Hyperscale operators take a lot of factors into account when deciding where to locate their datacenter infrastructure," said Synergy chief analyst John Dinsdale. "This includes availability of suitable real estate, cost and availability of power supply options, proximity to customers, the risk of natural disasters, local incentives and approvals processes, the ease of doing business and internal business dynamics, and this has inevitably led to some hyperscale hot spots." Amazon in particular locates a large amount of its datacenter infrastructure in Northern Virginia, with Microsoft, Facebook, Google, ByteDance, and others also having a major presence, according to Synergy. The big three cloud providers -- Amazon, Microsoft and Google -- have the broadest hyperscale bit barn footprint, with each of these having over 130 datacenters of the 800 or so around the globe. When measured in datacenter capacity, the leading companies are Amazon, Google, Microsoft, Facebook, Alibaba and Tencent, according to Synergy.

An anonymous reader quotes a report from New Scientist: A nuclear fusion reaction has lasted for 30 seconds at temperatures in excess of 100 million degrees celsius. While the duration and temperature alone aren't records, the simultaneous achievement of heat and stability brings us a step closer to a viable fusion reactor -- as long as the technique used can be scaled up. [...] Now Yong-Su Na at Seoul National University in South Korea and his colleagues have succeeded in running a reaction at the extremely high temperatures that will be required for a viable reactor, and keeping the hot, ionized state of matter that is created within the device stable for 30 seconds.

Controlling this so-called plasma is vital. If it touches the walls of the reactor, it rapidly cools, stifling the reaction and causing significant damage to the chamber that holds it. Researchers normally use various shapes of magnetic fields to contain the plasma -- some use an edge transport barrier (ETB), which sculpts plasma with a sharp cut-off in pressure near to the reactor wall, a state that stops heat and plasma escaping. Others use an internal transport barrier (ITB) that creates higher pressure nearer the center of the plasma. But both can create instability. Na's team used a modified ITB technique at the Korea Superconducting Tokamak Advanced Research (KSTAR) device, achieving a much lower plasma density. Their approach seems to boost temperatures at the core of the plasma and lower them at the edge, which will probably extend the lifespan of reactor components.

Dominic Power at Imperial College London says that to increase the energy produced by a reactor, you can make plasma really hot, make it really dense or increase confinement time. "This team is finding that the density confinement is actually a bit lower than traditional operating modes, which is not necessarily a bad thing, because it's compensated for by higher temperatures in the core," he says. "It's definitely exciting, but there's a big uncertainty about how well our understanding of the physics scales to larger devices. So something like ITER is going to be much bigger than KSTAR". Na says that low density was key, and that "fast" or more energetic ions at the core of the plasma -- so-called fast-ion-regulated enhancement (FIRE) -- are integral to stability. But the team doesn't yet fully understand the mechanisms involved. The reaction was stopped after 30 seconds only because of limitations with hardware, and longer periods should be possible in future. KSTAR has now shut down for upgrades, with carbon components on the wall of the reactor being replaced with tungsten, which Na says will improve the reproducibility of experiments. The research has been published in the journal Nature.

A timely mobile alert may have prevented hundreds of thousands of Californians from being plunged into darkness in the middle of a heat wave Tuesday night. Bloomberg reports: Just before 5:30 p.m. local time, California's grid operator ordered its highest level of emergency, warning that blackouts were imminent. Then, at 5:48 p.m., the state's Office of Emergency Services sent out a text alert to people in targeted counties, asking them to conserve power if they could. Within five minutes the grid emergency was all but over. Power demand plunged by 1.2 gigawatts between 5:50 and 5:55 p.m., and would continue to drop in the hours after that, according to data from the California Independent System Operator. A gigawatt is enough to power about 750,000 Californian homes.

But while the state's grid operator said California had avoided rolling blackouts Tuesday, some cities apparently didn't get the message. Officials in three San Francisco Bay area cities -- Alameda, Healdsburg and Palo Alto -- reported on social media that power shutdowns were underway that evening, which also could have contributed to the sharp decline in demand. By 8 p.m., the grid operator canceled the highest level of emergency without calling for power cuts. More than 500,000 homes and businesses had been warned earlier in the day that they might lose service.

Apple today announced that all iPhone 14 models sold in the U.S. do not have a built-in SIM card tray and instead rely entirely on eSIM technology. MacRumors reports: Tech specs on Apple's website confirm the iPhone 14, iPhone 14 Plus, iPhone 14 Pro, and iPhone 14 Pro Max are not compatible with physical SIM cards and instead have dual eSIM support, allowing for multiple cellular plans to be activated on a single device. An eSIM is a digital SIM that allows users to activate a cellular plan without having to use a physical nano-SIM card. eSIM availability is rapidly expanding, but the technology is still not available in all countries, which explains why iPhone 14 models will remain available with a SIM card tray outside of the U.S. for now. Apple's website has a list of carriers that support eSIM technology around the world. In the U.S., this includes AT&T, T-Mobile, US Cellular, Verizon, Xfinity Mobile, Boost Mobile, H2O Wireless, Straight Talk, C Spire, and some others.

Today, Apple announced a handful of new products at their iPhone launch event in Cupertino. Along with four new iPhones, Apple unveiled the brand-new Series 8 Apple Watch and refreshed SE. It also unveiled a completely new, rugged Apple Watch Ultra. The Verge reports: The Apple Watch Ultra is the star of the show because it isn't something we've seen before. It's got a big honking 49mm rectangular display, which... truly is in a class of its own. Not only is it the biggest Apple Watch screen, but it's also the brightest at 2,000 nits. As for how that wrist slab feels, it was actually lighter on my wrist than I'd expected, probably because its case is made of titanium. But make no mistake -- it is a BIG watch. Another thing that's immediately apparent is the design tweaks Apple's made for the extreme fitness crowd. There's the new orange action button, a button guard, and a redesigned crown. The rim around the display is also raised to protect the sapphire crystal display -- which is something we saw on the Samsung Galaxy Watch 5 Pro as well. Plus, that display is truly flat.

Next up is the Apple Watch Series 8. At a glance, there's really not too much to differentiate it from the Series 7, though I can definitely tell the screen appears bigger. Inside, there's a new temperature sensor, though it's not something I was able to really check out here in Cupertino. The sensor is meant to help retrospectively detect ovulation. It's also got an updated chip -- the S8 -- which keeps things snappy when you're swiping through menus. The Series 8 also has a new gyroscope and accelerometer to help call emergency services if it detects you've been in a car crash.

The new SE is also a fairly incremental update. It, too, gets an upgrade to the S8 chip, whereas the original SE was a bit of a Frankenstein watch. It had the processor of the Series 5 mixed in with some sensors used in the Series 6 minus the EKG sensor. The new SE has the same motion sensors as the Series 8 for crash detection and is 20 percent faster than before. The screen is also 30 percent larger than the Series 3, and the sensor array color matches the front. It's mostly still missing the always-on display and the new temperature sensor. All three models are available for preorder today. The Series 8 and SE will ship on September 16, with the Ultra shipping on the 23rd.

Beeftopia writes: Germany will keep two of its three reactors operational through the end of 2022, despite a pledge to shut them all down by December 31. A likely winter gas shortage prompted the change. "German Economy Minister Robert Habeck said in a statement on Monday the move did not mean Berlin was reneging on its long-standing promise to exit nuclear energy by the end of 2022," reports Reuters. Habeck went on to say, "It remains very improbable that we will have crisis situations and extreme scenarios" requiring further use of nuclear.

"Germany is part of a European system hit by a decline in Russian gas deliveries, the French nuclear power squeeze and a drought that has curbed hydroelectric production and cooling water supplies to thermal power stations as well as hampering barge deliveries of coal," Reuters says. A problem with the planned use case for the reactors is that nuclear plants are not designed to be variable backup energy generators, but rather continuous first-line generators.

Google has started sending out invites for its fall hardware event, which is set to take place on Thursday, October 6th, at 10AM ET. The Verge reports: The event will launch the upcoming Pixel 7 phones, as well as the Pixel Watch -- the company showed off both devices at its I/O event earlier this year, announcing they're coming in the fall. Since Google's last event, purported leaks have revealed even more about the devices, with both the Watch and phones seemingly making their way into people's hands.

The Pixel 7 seems like it'll be a pretty straightforward continuation of the previous generation of phones, but there's a bit of pressure with the Watch as the "first smartwatch built by Google, inside and out," as its product page on Google's online store says. Google's competitors (namely Samsung and Apple) have been making wearables for years, so the company's entry has to be competent and at least somewhat competitive right out of the gate. So far, details about it are scant -- we don't even know what chip it'll use -- but we should get a clearer picture come October 6th.

Leading foundry operators stand to benefit from a 25 percent investment tax credit (ITC) on domestic fab projects, according to a document published by the US Department of Commerce this week. The Register reports: The document [PDF] sheds new light on the department's plans for divvying up the more than $50 billion CHIPS Act fund approved this year by Congress to fund domestic chip production. The tax credits accompany $39 billion in grants, cooperative agreements, loans, and loan guarantees are available to companies working to advance US semiconductor and supply chain security interests.

The Commerce Department appears to have mixed sentiments about the state of the US in the global semiconductor sphere. The document argues the US remains the world leader in chip design and native design and automation tools but also notes the US is responsible for only 10 percent of global chip capacity and just 3 percent of global packaging, assembling, and testing services, pointing to areas where America has fallen behind when it comes to domestic production. The department adds that recent advancements made by the People's Republic of China to accelerate their own domestic chip manufacturing capacity has only served to exacerbate the risk to US supply chains.

With that said, the Commerce Department isn't being picky about which companies are eligible for funding. Any company, foreign or domestic, that takes steps to advance the commerce departments goals, with the exception of "entities of concern." Those goals in include accelerating leading-edge and legacy chip production in the US, research and development into next-generation semiconductor applications, and efforts to develop an adequate workforce to fuel this expansion. While foreign manufacturers aren't excluded from receiving funding, the Commerce Department emphasizes that those funds must go towards domestic infrastructure and can't be used abroad.

In 2021, 9% of the world's auto sales were electric vehicles, reports CNBC (citing statistics from the International Energy Agency). Yet CNBC also notes that electric vehicles "require 30% fewer parts and components manufacturing than conventional cars," according to researchers for an Industrial Heartland case study."

So will that create problems in America's heartland? "Large swaths of the Midwest have economies based around the auto parts manufacturing trade..." "When we look carefully at what goes on on the factory floor, it won't be less workers," Keith Cooley, former head of Michigan's Labor Department, told CNBC. "There will be different people building the cars." Researchers believe modern factory jobs will require more education and could be less available than they were in the past. They estimate that electric vehicles could require 30% less manufacturing labor when compared with conventional cars. "The lines that run to drive oil or gas around an internal combustion engine aren't going to be there," said Cooley.

This change could hit the parts suppliers in the auto industry.

A surprise from the Associated Press. The head of America's Environment Protection Agency "said Friday that advanced nuclear technology will be 'critical' for both the United States and Japan as they step up cooperation to meet decarbonization goals." Michael Regan, after holding talks with his Japanese counterpart Akihiro Nishimura in Tokyo, told a joint news conference that nuclear energy in their countries plays a role... "I think the science tells us that we have to respond to the climate crisis with a sense of urgency and nuclear energy and nuclear technology has and can have a role in continuing with a zero emissions contribution to the climate," he said, showing support for Japan's recent shift toward returning to nuclear energy.

Japanese Prime Minister Fumio Kishida said last week he instructed his government to consider developing safer, smaller nuclear reactors, in a renewed emphasis on nuclear energy years after many of the country's plants were shut down. Kishida said Japan needs to consider all options of energy mix, including nuclear, to bolster its "green transformation" effort to curb emissions of greenhouse gases and to secure stable energy supply. Japan has pledged to reach carbon neutrality by 2050.... While maintaining a 20%-22% target for nuclear energy as part of its energy mix for 2030, Japan's government had previously insisted it was not considering building new plants or replacing aged reactors, apparently to avoid stoking criticism from a wary public. Kishida's recent comment represents a sharp change from that stance.

Kishida asked a government panel to decide by the end of the year on its proposal for the development and construction of "new innovative reactors," such as small modular reactors, while also considering extending operational lifespan of aging reactors.

President Joe Biden also believes that it is an "all hands on deck approach," Regan said, adding that investments in nuclear technology and carbon capture will complement the big potential for renewable energy.

Sunnova is one of America's largest rooftop solar companies, according to the New York Times. But they've now applied to California's Public Utilities Commission for permission to become the state's first solar (and storage) micro-utility, initiating formal steps to qualify and "request a certificate to construct and operate microgrids," targetting new home developments that aren't yet connected to the grid.

"We see a future where communities, neighborhoods, and businesses can operate independently from the legacy grid with sustainable energy sources that provide uninterrupted power," says the company's founder and CEO. "We believe microgrids address a strong need in the market for more robust energy solutions and better connectivity...." But he's also offering touting another possible benefit: "relief that the existing transmission and distribution system will experience given that most of the power that will be consumed by these communities will be generated locally from renewable resources."

The company likes to point out that America's recently-passed climate bill included tax incentives to encourage microgrids. But the New York Times describes it as "a business model that is illegal in much of the United States." Sunnova said it would offer those residents electricity that was up to 20 percent cheaper than the rates charged by investor-owned utilities like Pacific Gas & Electric and Southern California Edison. If approved by regulators, the micro-utility model, also known as a microgrid, could undermine the growth of those larger utilities by depriving them access to new homes or forcing them to lower their rates to keep that business. Sunnova executives argue that the approach they are seeking approval for was authorized under a California law passed almost two decades ago for a resort just south of Lake Tahoe. In addition, the company says advances in solar and battery technology mean that neighborhoods can be designed to generate more than enough electricity to meet their own needs at a lower cost than relying on the grid.

"If they don't want to choose me, that should be their right; if they don't want to choose you, that should be their right, too," said John Berger, the chief executive of Sunnova.

A small number of homeowners have gone off the grid as the cost of solar panels and batteries has fallen. But doing so can be hard or impossible. Some local governments have rejected permits for off-grid homes on health and safety grounds, arguing that a connection to the grid is essential. But connecting a single home to the grid can cost tens or even hundreds of thousands of dollars, which means an off-grid system may actually be cheaper — especially for properties in remote areas, or in places where the local grid is at its capacity and would require significant upgrades to serve more homes. Off-grid setups can also be appealing because once a system is paid off, the cost of operating and maintaining it is often modest and predictable, whereas utility rates can move up sharply.... The nationwide average retail electricity rate increased 11 percent in June from a year earlier, according to the Energy Information Administration.

But the kind of micro-utilities that Sunnova hopes to create have also had problems. The utopian visions of generating electricity where it is used have often run into maintenance and other problems. Many tiny utilities created under such models in the United States and Canada were later swallowed up by larger power companies.... Sunnova's microgrid approach could suffer a similar fate. But the costs of solar panels and batteries have tumbled over the last decade, making the energy that off-grid systems generate much more affordable....

Utilities have been pressing regulators to reduce the compensation homeowners receive for the excess solar energy their rooftop systems send to the grid. The companies have argued that customers with solar panels are being offered generous credits for power that they are not contributing adequately toward the cost of maintaining power lines and other grid equipment....

Building and operating microgrids could provide a steady source of income to companies like Sunnova. That could essentially transform the rooftop solar companies into the kinds of utilities that they have long fought against.
Sunnova bills itself as an "Energy as a Service" company, and they expect their microgrids to experience 30 minutes or less of outages each year, the Times points out, "compared with an average of two hours a year at California's large investor-owned utilities."

In the article, the chief executive of home-building company Lennar says they've already formed a partnership with Sunnova. "We value the current electric grid and we're intrigued by new microgrid solutions that can supplement and support the traditional utility grid and help solve reliability during extreme weather and peak demand."

Long-time Slashdot reader swell shared Scientific American's report on the quest for neuromorphic chips: The human brain is an amazing computing machine. Weighing only three pounds or so, it can process information a thousand times faster than the fastest supercomputer, store a thousand times more information than a powerful laptop, and do it all using no more energy than a 20-watt lightbulb. Researchers are trying to replicate this success using soft, flexible organic materials that can operate like biological neurons and someday might even be able to interconnect with them. Eventually, soft "neuromorphic" computer chips could be implanted directly into the brain, allowing people to control an artificial arm or a computer monitor simply by thinking about it.

Like real neurons — but unlike conventional computer chips — these new devices can send and receive both chemical and electrical signals. "Your brain works with chemicals, with neurotransmitters like dopamine and serotonin. Our materials are able to interact electrochemically with them," says Alberto Salleo, a materials scientist at Stanford University who wrote about the potential for organic neuromorphic devices in the 2021 Annual Review of Materials Research. Salleo and other researchers have created electronic devices using these soft organic materials that can act like transistors (which amplify and switch electrical signals) and memory cells (which store information) and other basic electronic components.

The work grows out of an increasing interest in neuromorphic computer circuits that mimic how human neural connections, or synapses, work. These circuits, whether made of silicon, metal or organic materials, work less like those in digital computers and more like the networks of neurons in the human brain.... An individual neuron receives signals from many other neurons, and all these signals together add up to affect the electrical state of the receiving neuron. In effect, each neuron serves as both a calculating device — integrating the value of all the signals it has received — and a memory device: storing the value of all of those combined signals as an infinitely variable analog value, rather than the zero-or-one of digital computers.

The European Commission is advocating new rules for mobile phone and tablet repairability. PC Magazine reports: Draft proposals published this week would require manufacturers to make at least 15 components available to professional repairers for up to five years after releasing a new phone in the European Union (EU). That means customers would get guaranteed access to replacement batteries, back covers, front- and rear-facing cameras, audio connectors, charging ports, microphones and speakers, SIM and memory card trays, and more.

"The steep increase in the demand for smartphones and tablets, combined [with] their increased functionality, has resulted in increased demand for energy and materials needed to manufacture these devices on the EU market, accompanied by an increase in their associated environmental impacts," Commission President Ursula Von Der Leyen wrote in the proposal. "In addition, devices are often replaced prematurely by users and are, at the end of their useful life, not sufficiently reused or recycled, leading to a waste of resources."

If adopted, the initiative would also usher in a new energy label for phones and tablets -- similar to the ones already in place across Europe for TVs and large household items. The labels would indicate an expected battery life, and include details on water and dust protection, and rate the device's resistance to drops and scratches. Those manufacturers, meanwhile, that can't (or won't) supply batteries for five years must instead meet a set of battery endurance tests that certify devices achieve 80% of a rated capacity after 1,000 full-charge cycles. They'll also need to ensure software updates never negatively impact battery life.

Lenovo is finally selling AR glasses for consumers with the recently-announced Lenovo Glasses T1, which Ars Technica's Scharon Harding got to demo. Here's an excerpt from her report: With their Micro OLED displays and required tether to Windows, macOS, Android, or iOS devices, they bring some notable features to a space that has piqued industry-wide interest but is still likely far from becoming ubiquitous. The early version of the T1 I tried had limited features; I was mostly only able to view a homepage with basic menu options and a desktop with icons for apps, like web browsing. Although the glasses weren't ready for me to watch a movie or hop around apps, I was impressed at how clear text and menu items were. This was in a sunny room with exceedingly tall windows. Even when facing sunlight, the few colors on display seemed vibrant and the text legible.

Lenovo specs the displays with 10,000:1 contrast and 1920x1080 pixels per eye. The glasses are also TUV-certified for low blue light and flicker reduction, according to Lenovo. Much more time is needed to explore and challenge the Micro OLED displays before I pass final judgment. But the combination of smaller pixels and, from what I saw thus far, strong colors, should accommodate screens so close to the eyes. More broadly speaking, brightness can be a concern with OLED technologies, but the small demo I saw fared well in a sun-flushed room.

I used the Glasses T1 while it was connected to an Android smartphone via its USB-C cable, but it's also supposed to work with PCs, macOS devices, and, via an adapter sold separately, iPhones. [...] With no processor or battery, it's easier for the glasses to stay trim. There are also no sensors or cameras like the Lenovo ThinkReality A3, announced last year, has. Other T1 features include a pair of speakers (one near each temple) and the ability to add prescription lenses. [...] The Glasses T1 are expected to be available in select markets in 2023 after debuting in China (as the Lenovo Yoga Glasses) this year. Lenovo didn't set a price, but I was told it's hoping to keep the glasses under $500.

Intel has made available versions of its 12th-generation Core processors optimized for edge and IoT applications, claiming the purpose-built chips enable smaller form factor designs, but with the AI inferencing performance to analyze data right at the edge. The Register reports: The latest members of the Alder Lake family, the 12th Gen Intel Core SoC processors for IoT edge (formerly Alder Lake PS) combine the performance profile and power envelope of the mobile chips but the LGA socket flexibility of the desktop chips, according to Intel, meaning they can be mounted directly on a system board or in a socket for easy replacement. Delivered as a multi-chip package, the new processors combine the Alder Lake cores with an integrated Platform Controller Hub (PCH) providing I/O functions and integrated Iris Xe graphics with up to 96 graphics execution units. [...]

Intel VP and general manager of the Network and Edge Compute Division Jeni Panhorst said in a statement that the new processors were designed for a wide range of vertical industries. "As the digitization of business processes continues to accelerate, the amount of data created at the edge and the need for it to be processed and analyzed locally continues to explode," she said. Another key capability for managing systems deployed in edge scenarios is that these processors include Intel vPro features, which include remote management capabilities built into the hardware at the silicon level, so an IT admin can reach into a system and perform actions such as changing settings, applying patches or rebooting the platform.

The chips support up to eight PCIe 4.0 lanes, and four Thunderbolt 4/USB4 lanes, with up to 64GB of DDR5 or DDR4 memory, and the graphics are slated to deliver four 4K displays or one 8K display. Operating system support includes Windows 10 IoT Enterprise 2021 Long Term Servicing Channel (LTSC) and Linux options. Intel said the new SoCs are aimed at a broad range of industries, including point-of-sale kit in the retail, banking, and hospitality sectors, industrial PCs and controllers for the manufacturing industry, plus healthcare.

A company called Subsea Cloud is planning to have a commercially available undersea datacenter operating off the coast of the US before the end of 2022, with other deployments planned for the Gulf of Mexico and the North Sea. The Register reports: Subsea, which says it has already deployed its technology with "a friendly government faction," plans to put its first commercial pod into the water before the end of this year near Port Angeles, Washington. The company claims that placing its datacenter modules underwater can reduce power consumption and carbon dioxide emissions by 40 percent, as well as lowering latency by allowing the datacenter to be located closer to metropolitan areas, many of which are located near the coast. However, according to Subsea founder Maxie Reynolds, it can also deploy 1MW of capacity for as much as 90 percent less cost than it takes to get 1MW up and running at a land-based facility.

The Port Angeles deployment, known as Jules Verne, will comprise one 20ft pod, which is similar in size and dimensions to a standard 20-foot shipping container (a TEU or Twenty-foot Equivalent Unit). Inside, there is space for about 16 datacenter racks accommodating about 800 servers, according to Subsea. Additional capacity, if and when required, is delivered by adding another pod. The pod-to-shore link in this deployment provides a 100Gbps connection. As it is a commercial deployment, Jules Verne will be open for any prospective clients or partners to come and check it out, virtually or otherwise, according to Reynolds. It will be sited in shallow water, visible from the port, whereas the Njord01 pod in the Gulf of Mexico and the Manannan pod in the North Sea are expected to be deeper, at 700-900ft and 600-700ft respectively.

The Subsea pods are kept cool by being immersed in water, which is one reason for the reduced power and CO2 emissions. Inside, the servers are also immersed in a dielectric coolant, which conducts heat but not electricity. However, the Subsea pods are designed to passively disperse the heat, rather than using pumps as is typical in submersion cooling in land-based datacenters. But what happens if something goes wrong, or a customer wants to replace their servers? According to Subsea, customers can schedule periodic maintenance, including server replacement, and the company says that would take 4-16 hours for a team to get to the site, bring up the required pod(s), and replace any equipment.

An anonymous reader quotes a report from CNBC: On the afternoon of April 14, 2020, dozens of firefighters arrived at an Amazon warehouse in Fresno, California, as thick plumes of smoke poured from the roof of the 880,000-square-foot warehouse. Some 220 solar panels and other equipment at the facility, known as FAT1, were damaged by the three-alarm fire, which was caused by "an undetermined electrical event within the solar system mounted on top of the roof," Leland Wilding, Fresno's fire investigator, wrote in an incident report. A little over a year later, about 60 firefighters were called to an even larger Amazon facility in Perryville, Maryland, to put out a two-alarm blaze, local news outlets reported. In the intervening months, at least four other Amazon fulfillment centers caught fire or experienced electrical explosions due to failures with their solar energy-generating systems, according to internal company documents viewed by CNBC.

The documents, which have never been made public, indicate that between April 2020 and June 2021, Amazon experienced "critical fire or arc flash events" in at least six of its 47 North American sites with solar installations, effecting 12.7% of such facilities. Arc flashes are a kind of electrical explosion. "The rate of dangerous incidents is unacceptable, and above industry averages," an Amazon employee wrote in one of the internal reports. [...] By June of last year, all of Amazon's U.S. operations with solar had to be taken offline temporarily, internal documents show. The company had to ensure its systems were designed, installed and maintained properly before "re-energizing" any of them.

Amazon spokesperson Erika Howard told CNBC in a statement that the incidents involved systems run by partners, and that the company responded by voluntarily turning off its solar-powered roofs. "Out of an abundance of caution, following a small number of isolated incidents with onsite solar systems owned and operated by third parties, Amazon proactively powered off our onsite solar installations in North America, and took immediate steps to re-inspect each installation by a leading solar technical expert firm," the statement said. [...] "As inspections are completed, our onsite solar systems are being powered back on," Howard said. "Amazon also built a team of dedicated solar experts overseeing the construction, operations, and maintenance of our systems in-house to ensure the safety of our systems." "An Amazon employee estimated, in the documents circulated internally, that each incident cost the company an average of $2.7 million," adds the report. "The Amazon employee also said the company would lose $940,000 per month, or $20,000 for each of the 47 decommissioned North American sites, as long as the solar remained offline. There could be additional costs for Amazon depending on contracts with clean energy partners for renewable energy credits, the documents show."