In a clear signal of its ambitions for the estimated $91.18 billion AI chip market, Intel this morning announced that it has acquired Habana Labs, an Israel-based developer of programmable AI and machine learning accelerators for cloud data centers. From a report: The deal is worth approximately $2 billion, and Intel says it'll strengthen its AI strategy as Habana begins to sample its proprietary silicon to customers. Habana -- which previously raised $75 million in venture capital last November -- will remain an independent business unit and will continue to be led by its current management team, and it'll report to Intel's data platforms group. Chairman Avigdor Willenz will serve as senior adviser to the business unit as well as to Intel.

"This acquisition advances our AI strategy, which is to provide customers with solutions to fit every performance need -- from the intelligent edge to the data center," said executive vice president and general manager of the data platforms group at Intel Navin Shenoy. "More specifically, Habana turbo-charges our AI offerings for the data center with a high-performance training processor family and a standards-based programming environment to address evolving AI [compute requirements]." Habana offers two silicon products targeting workloads in AI and machine learning: the Gaudi AI Training Processor and the Goya AI Inference Processor.

An anonymous reader quotes Mike Melanson's "This Week in Programming" column: WebAssembly is a binary instruction format for a stack-based virtual machine and this week, the World Wide Web Consortium (W3C) dubbed it an official web standard and the fourth language for the Web that allows code to run in the browser, joining HTML, CSS and JavaScript... With this week's news, WebAssembly has officially reached version 1.0 and is supported in the browser engines for Firefox, Chrome, Safari, and Internet Explorer, and the Bytecode Alliance launched last month to help ensure "a WebAssembly ecosystem that is secure by default" and for bringing WebAssembly to outside-the-browser use.

Of course, not everything is 100% rosy. As pointed out by an article in The Register, WebAssembly also brings with it an increased level of obfuscation of what exactly is going on, giving it an increased ability to perform some surreptitious actions. For example, they cite one study that "found 'over 50 percent of all sites using WebAssembly apply it for malicious deeds, such as [crypto] mining and obfuscation.'" Nonetheless, with WebAssembly gaining this designation by W3C, it is, indeed, time to pay closer attention to the newly nominated Web language standard.

At least one major provider of hardware-level BIOS drivers is actively deleting old stuff it no longer supports, while old FTP sites where vintage drivers are often found are soon going to be harder to reach. Ernie Smith, writing for Motherboard: You've never lived until you've had to download a driver from an archived forum post on the Internet Archive's Wayback Machine. You have no idea if it's going to work, but it's your only option. So you bite the bullet. I recently did this with a PCI-based SATA card I was attempting to flash to support a PowerPC-based Mac, and while it was a bit of a leap of faith, it actually ended up working. Score one for chance. But this, increasingly, feels like it may be a way of life for people trying to keep old hardware alive -- despite the fact that all the drivers generally have to do is simply sit on the internet, available when they're necessary.

Apparently, that isn't easy enough for Intel. Recently, the chipmaker took BIOS drivers, a boot-level firmware technology used for hardware initialization in earlier generations of PCs, for a number of its unsupported motherboards off its website, citing the fact that the programs have reached an "End of Life" status. While it reflects the fact that Unified Extensible Firmware Interface (UEFI), a later generation of firmware technology used in PCs and Macs, is expected to ultimately replace BIOS entirely, it also leaves lots of users with old gadgets out in a lurch. And as Bleeping Computer has noted, it appears to be part of a broader trend to prevent downloads for unsupported hardware on the Intel website -- things that have long lived past their current lives. After all, if something goes wrong, Intel can be sure it's not liable if a 15-year-old BIOS update borks a system.

An anonymous reader quotes a report from ZDNet: Academics from three universities across Europe have disclosed today a new attack that impacts the integrity of data stored inside Intel SGX, a highly-secured area of Intel CPUs. The attack, which researchers have named Plundervolt, exploits the interface through which an operating system can control an Intel processor's voltage and frequency -- the same interface that allows gamers to overclock their CPUs. Academics say they discovered that by tinkering with the amount of voltage and frequency a CPU receives, they can alter bits inside SGX to cause errors that can be exploited at a later point after the data has left the security of the SGX enclave. They say Plundervolt can be used to recover encryption keys or introduce bugs in previously secure software. Intel desktop, server, and mobile CPUs are impacted. A full list of vulnerable CPUs is available here. Intel has also released microcode (CPU firmware) and BIOS updates today that address the Plundervolt attack [by allowing users to disable the energy management interface at the source of the attack, if not needed]. Proof-of-concept code for reproducing attacks will be released on GitHub.

Intel chief executive Bob Swan says he's willing to let go the company's traditional dominance of the market for CPUs in order to meet the rising demand for newer, more specialized silicon chips for applications such as AI and autonomous cars. From a report: Intel's Bob Swan blames being focused on 90% CPU market share as a reason for missing opportunities and transitions, envisions Intel as having 30% of all-silicon TAM instead of majority CPU TAM. Just a few years ago, Intel owned more than 90% of the market share in the x86 CPU market. Many financial models used Intel's revenue as a proxy for the Total Available Market of the CPU sector. With a full-year revenue of $59.4 billion in 2017, you can estimate the total TAM of the CPU side of things at roughly $66 billion (2017 est). Bob Swan believes that this mindset of protecting a majority share in the CPU side has led to Intel becoming complacent and missing out on major opportunities. Bob even went as far as to say that he is trying to "destroy" this thinking of having a 90% market share in the CPU side and instead wants people to come into office thinking Intel has 30% market share in "all Silicon." Swan on how Intel got to the place where it is now: How we got here is really kind of threefold, one we got a lot faster than we expected and the demand for CPUs and servers grew much faster than we expected in 2018. You'll remember we came into 2018 projecting a 10% growth and we grew by 21% growth so the good news problem is that demand for our products in our transformation to a data-centric company was much higher than we expected. Secondly, we took on a 100% market share for smartphone modem and we decided that we would build it in our fabs, so we took on even more demand. And third, to exacerbate that, we slipped on bringing our 10nm to life and when that happens you build more and more performance into your last generation for us -- 14nm -- which means there is a higher core count and larger die size. So those three -- growing much faster than we thought, bringing modems inside and delaying 10nm resulted in a position where we didn't have flexible capacity.

China has ordered that all foreign computer equipment and software be removed from government offices and public institutions within three years, the Financial Times reports. hackingbear writes: The government directive is likely to be a blow to US multinational companies like HP, Dell and Microsoft and mirrors attempts by Washington to limit the use of Chinese technology, as the trade war between the countries turns into a tech cold war. The Trump administration banned US companies from doing business with Chinese Chinese telecommunications company Huawei earlier this year and in May, Google, Intel and Qualcomm announced they would freeze cooperation with Huawei. By excluding China from western know-how, the Trump administration has made it clear that the real battle is about which of the two economic superpowers has the technological edge for the next two decades. This is the first known public directive from Beijing setting specific targets limiting China's use of foreign technology, though it is part a wider move within China to increase its reliance on domestic technology.

An anonymous reader quotes a report from The Wall Street Journal: Huawei's latest phone, which it unveiled in September -- the Mate 30 with a curved display and wide-angle cameras that competes with Apple's iPhone 11 -- contained no U.S. parts, according to an analysis by UBS and Fomalhaut Techno Solutions, a Japanese technology lab that took the device apart to inspect its insides. In May, the Trump administration banned U.S. shipments to Huawei as trade tensions with Beijing escalated. That move stopped companies like Qualcomm Inc. and Intel Corp. from exporting chips to the company, though some shipments of parts resumed over the summer after companies determined they weren't affected by the ban.

While Huawei hasn't stopped using American chips entirely, it has reduced its reliance on U.S. suppliers or eliminated U.S. chips in phones launched since May (Warning: source paywalled; alternative source), including the company's Y9 Prime and Mate smartphones, according to Fomalhaut's teardown analysis. Similar inspections by iFixit and Tech Insights Inc., two other firms that take apart phones to inspect components, have come to similar conclusions. With the Mate 30, audio chips supplied in older versions came from Cirrus Logic. In the newer Mate 30 models, chips were provided by NXP Semiconductors NV, a Dutch chip maker, according to Fomalhaut. Power amplifiers provided by Qorvo or Skyworks were replaced with chips from HiSilicon, Huawei's in-house chip design firm, the teardown analysis showed. A Huawei spokesman said it is the company's "clear preference to continue to integrate and buy components from U.S. supply partners. If that proves impossible because of the decisions of the U.S. government, we will have no choice but to find alternative supply from non-U.S. sources."

An anonymous reader quotes a report from Bloomberg: Amazon Web Services has developed a more powerful version of its own chips to power services for cloud-computing customers, as well as some of AWS's own programs. AWS Chief Executive Andy Jassy on Tuesday introduced a second-generation chip, called Graviton2, aimed at general-purpose computing tasks. He didn't specify a release date. The company last year unveiled its first line of Graviton chips, which it said would support new versions of its main EC2 cloud-computing service. Prior to that, Amazon -- and other big cloud operators -- had almost exclusively used Intel Xeon chips. The company said at the time that the Graviton-backed cloud service would be available at a "significantly lower cost" than existing offerings run on Intel processors. Jassy said Intel is "a very close partner," but to push the envelope on prices, "we had to do some innovating ourselves."

"Amazon is using its 2015 acquisition of startup Annapurna Labs, which Jassy called a 'a big turning point for us,' to design its own chips," reports Bloomberg. "The new processor uses technology from SoftBank Group Corp. unit ARM Holdings, a standard that dominates in mobile phones."

Intel sold its smartphone modem chip business to Apple at "a multi-billion dollar loss," the U.S. chipmaker said in a court filing on Friday, alleging that rival Qualcomm forced it out of the market. From a report: Intel made the claims in a brief filed with the 9th U.S. Circuit Court of Appeals, where Qualcomm is seeking to overturn a sweeping antitrust decision against it after losing a lawsuit by the U.S. Federal Trade Commission. Intel, whose executives testified at the trial, argued on Friday that the ruling should stand. Appeal proceedings are expected to begin in January. In a 233-page decision issued in May, U.S. District Judge Lucy Koh in San Jose wrote that Qualcomm's patent licensing practices "strangled competition" in parts of the market for modem chips that connect smartphones to mobile data networks. She ordered the San Diego-based company to renegotiate licensing agreements at reasonable prices.

Amazon's cloud computing unit has designed a second, more powerful generation of datacenter processor chip, Reuters reported Thursday, the latest sign that the e-commerce company is pouring money into custom silicon for its fastest-growing business. From a report: The new Amazon Web Services chip uses technology from SoftBank Group-owned Arm Holdings, the sources said. One of the sources familiar with the matter said it will be at least 20% faster than Amazon's first Arm-based chip, named Graviton, which was released last year as a low-cost option for easier computing tasks. If Amazon Web Services' chip efforts are successful, it could lessen the unit's reliance on Intel and Advanced Micro Devices for server chips. In cloud computing, businesses rent out servers from Amazon instead of running their own datacenters. Analysts expect Amazon's cloud unit to generate $34.9 billion in sales in 2019, according to IBES data from Refinitiv.

An anonymous reader quotes a report from Ars Technica: In April of this year, Intel cancelled its 5G-modem building plans. This week, it's announcing that they're back on the table -- but this time, with system-on-chip vendor MediaTek building the hardware. The partnership has Intel setting the 5G specifications, MediaTek developing the modem to match, and Intel optimizing and validating it afterwards. Intel will also lend its marketing and integration muscle to convince OEMs to use the new hardware and help them make sure it works well in final products. This also means Intel will be writing operating-system-level drivers for the modems.

The partnership looks like a sensible one for both parties: Intel has been struggling to get its own 10nm hardware out the door on time, so getting this hardware design task off its plate may relieve some pressure there, while still keeping the company in an emerging market. MediaTek, on the other hand, can definitely benefit from Intel's software development expertise and deep integration with OEM vendors in the PC space. Specifically, the companies will be adapting MediaTek's existing Helio M70 5G modem for use in PC hardware. The M70 modem is already being built into MediaTek's Dimensity family of ARM System-on-Chip (SoC) designs; the new partnership gives MediaTek a whole new platform to market to and gives Intel a foot back into the door in 5G. It also may represent a way for Intel to push back against ARM-based Windows hardware like Samsung's Galaxy Book S, built on Qualcomm's Snapdragon 8cx platform. We can expect to see the resulting hardware shipping some time in 2021, the report adds.

MojoKid writes: Intel and AMD have been duking it out in the high-end desktop processor space lately. AMD's return to competitive footing versus Intel has propelled the company forward and the brand has a loyal, passionate following due the competitive performance-per-dollar its 3rd Generation Ryzen processors bring versus Intel offerings. Today, both companies have launched new flagship many-core CPUs, the Intel Core i9-10980XE, which is an 18-core chip, and the AMD 3rd Gen Threadripper 3970X and 3960X, which are 32-core and 24-core chips, respectively. Intel's Core i9-10980XE brings a lower price of $999 and competes more favorably versus AMD's lower-end 16-core Ryzen 9 3950X that's priced at just $750. Meanwhile, the new AMD Threadripper 3960X at $1399 and Threadripper 3970X at $1999 leave Intel's fastest desktop chip in the dust in multi-threaded workloads, sometimes by a wide margin. In addition, while Threadripper 3960X and 3970X pull only about 26 to 36 Watts of additional power versus Intel's new Core i9-10980XE, they do it with 33-77% more core resources. Regardless, it's impressive how the tables have turned, as AMD is now firmly entrenched with some better value propositions in high-end desktop processors, and better performance in many cases as well.

diegocg writes: Linux 5.4 has been released, featuring the new kernel lockdown mode, intended to strengthen the boundary between UID 0 and the kernel; virtio-fs, a high-performance virtio driver which allows a virtualized guest to mount a directory that has been exported on the host; fs-verity, for detecting file tampering, like dm-verity, but works on files rather than block devices; dm-clone, which allows live cloning of dm targets; two new madvise() flags for improved app memory management on Android, support for new Intel/AMD GPUs, support for the exfat file system and removing the experimental status of the erofs file system; a new haltpoll cpuidle driver and governor that greatly improves performance for virtualized guests wanting to do guest-side polling in the idle loop; and blk-iocost, a new cgroup controller that attempts to calculate more accurately the cost of IO. As always, many other new drivers and improvements can be found in the changelog.

Taiwan's MediaTek has announced a partnership with U.S. chipmaking giant Intel to supply future Intel-powered PCs with fifth-generation networking modems from the start of 2021. From a report: The agreement marks a small step toward a big change in the way computing is done, as 5G promises to revolutionize both the speed and availability of cellular networks, creating dense coverage with bandwidth comparable to current Wi-Fi standards and beyond. Mobile computers stand to benefit greatly from this upgrade, and U.S. PC vendors Dell and HP have both been named by MediaTek among the likely first customers for the 5G-enabled, Intel-powered laptops that are to come. In July, Intel agreed to sell its cellular modem business to Apple for $1 billion, which the Cupertino, California company will use to speed up and improve design efforts around a 5G chip for its 2020 iPhones.

MojoKid writes: Intel has unveiled its first discrete GPU solution that will hit the market in 2020, code name Ponte Vecchio. Based on 7nm silicon manufacturing and stack chiplet design with Intel's Foveros tech, Ponte Vecchio will target HPC markets for supercomputers and AI training in the datacenter. According to HotHardware, Ponte Vecchio will employ a combination of both its Foveros 3D packaging and EMIB (Embedded Multi-die Interconnect Bridge) technologies, along with High Bandwidth Memory (HBM) and Compute Express Link (CXL), which will operate over the newly ratified PCIe 5.0 interface and serve as Ponte Vecchio's high-speed switch fabric connecting all GPU resources. Intel is billing Ponte Vecchio as its first exascale GPU, proving its meddle in the U.S. Department of Energy's (DOE) Aurora supercomputer. Aurora will employ a topology of six Ponte Vecchio GPUs and two Intel Xeon Scalable processors based on Intel's next generation Sapphire Rapids architecture, along with Optane DC Persistent Memory on a single blade. The new supercomputer is schedule to arrive sometime in 2021.

An anonymous reader quotes a report from ZDNet: In the latest Top 500 supercomputer ratings, the average speed of these Linux-powered racers is now an astonishing 1.14 petaflops. The fastest of the fast machines haven't changed since the June 2019 Top 500 supercomputer list. Leading the way is Oak Ridge National Laboratory's Summit system, which holds top honors with an HPL result of 148.6 petaflops. This is an IBM-built supercomputer using Power9 CPUs and NVIDIA Tesla V100 GPUs. In a rather distant second place is another IBM machine: Lawrence Livermore National Laboratory's Sierra system. It uses the same chips, but it "only" hit a speed of 94.6 petaflops.

Close behind at No. 3 is the Sunway TaihuLight supercomputer, with an HPL mark of 93.0 petaflops. TaihuLight was developed by China's National Research Center of Parallel Computer Engineering and Technology (NRCPC) and is installed at the National Supercomputing Center in Wuxi. It is powered exclusively by Sunway's SW26010 processors. Sunway's followed by the Tianhe-2A (Milky Way-2A). This is a system developed by China's National University of Defense Technology (NUDT). It's deployed at the National Supercomputer Center in China. Powered by Intel Xeon CPUs and Matrix-2000 accelerators, it has a top speed of 61.4 petaflops. Coming at No. 5, the Dell-built, Frontera, a Dell C6420 system is powered by Intel Xeon Platinum processors. It speeds along at 23.5 petaflops. It lives at the Texas Advanced Computing Center of the University of Texas. The most powerful new supercomputer on the list is Rensselaer Polytechnic Institute Center for Computational Innovations (CCI)'s AiMOS. It made the list in the 25th position with 8.0 petaflops. The IBM-built system, like Summit and Sierra, is powered by Power9 CPUs and NVIDIA V100 GPUs. In closing, ZDNet's Steven J. Vaughan-Nichols writes: "Regardless of the hardware, all 500 of the world's fastest supercomputers have one thing in common: They all run Linux."

By Friday this week, Intel plans to remove old drivers and BIOS updates from its official website. From a report: "This download, BIOS Update [BLH6710H.86A] 0163, will no longer be available after November 22, 2019 and will not be supported with any additional functional, security, or other updates," reads a message posted to the download page of one of the impacted components. "Intel recommends that users of BIOS Update [BLH6710H.86A] 0163 uninstall and/or discontinue use as soon as possible," the message continues. The downloads are drivers and BIOS updates for Intel desktop components and motherboards the company released in the 90s and early-to-mid 2000s. Downloads for hundreds of components are believed to have been impacted, from motherboards to NIC cards and graphics cards. Most of the drivers are for Windows versions like 98, ME, XP, and older Windows Server editions -- old Windows OS versions that have themselves reached end-of-life (EOL) All components and motherboards reached (EOL) years ago, and Intel stopped delivering firmware updates as a result. Its website was merely hosting the older files for convenience.

MojoKid writes: AMD officially launched its latest many-core Zen 2-based processor today, a 16-core/32-thread beast known as the Ryzen 9 3950X. The Ryzen 9 3950X goes head-to-head against Intel's HEDT flagship line-up like the 18-core Core i9-9980XE but at a much more reasonable price point of $750 (versus over $2K for the Intel chip). The Ryzen 9 3950X has base and boost clocks of 3.5GHz and 4.7GHz, respectively. The CPU cores at the heart of the Ryzen 9 39050X are grouped into two, 7nm 8-core chiplets, each with dual, four-core compute complexes (CCX). Those chiplets link to an IO die that houses the memory controller, PCI Express lanes, and other off-chip IO. The new 16-core Zen 2 chips also use the same AM4 socket and are compatible with the same motherboards, memory, and coolers currently on the market for lower core-count AMD Ryzen CPUs. Throughout all of Hot Hardware's benchmark testing, the 16-core Ryzen 9 3950X consistently finished at or very near the top of the charts in every heavily-threaded workload, and handily took Intel's 18-core chip to task, beating it more often than not.

Both Microsoft and the Linux kernel teams have added ways to disable support for Intel Transactional Synchronization Extensions (TSX). From a report: TSX is the Intel technology that opens the company's CPUs to attacks via the Zombieload v2 vulnerability. Zombieload v2 is the codename of a vulnerability that allows malware or a malicious threat actor to extract information processed inside a CPU, information to which they normally shouldn't be able to access due to the security walls present inside modern-day CPUs. This new vulnerability was disclosed earlier this week. Intel said it would release microcode (CPU firmware) updates -- available on the company's Support & Downloads center. But, the reality of a real-world production environment is that performance matters. Past microcode updates for other attacks, such as Meltdown, Spectre, Foreshadow, Fallout, and Zombieload v1, have been known to introduce performance hits of up to 40%. Seeing that all the CPU attacks listed above are not only theoretical but also hard to pull off, some companies don't see this performance hit as an option.

An anonymous reader writes: A team of academics has disclosed today two vulnerabilities known collectively as TPM-FAIL that could allow an attacker to retrieve cryptographic keys stored inside TPMs. The first vulnerability is CVE-2019-11090 and impacts Intel's Platform Trust Technology (PTT). Intel PTT is Intel's fTPM software-based TPM solution and is widely used on servers, desktops, and laptops, being supported on all Intel CPUs released since 2013, starting with the Haswell generation. The second is CVE-2019-16863 and impacts the ST33 TPM chip made by STMicroelectronics. This chip is incredibly popular and is used on a wide array of devices ranging from networking equipment to cloud servers, being one of the few chips that received a CommonCriteria (CC) EAL 4+ classification — which implies it comes with built-in protection against side-channel attacks like the ones discovered by the research team. Unlike most TPM attacks, these ones were deemed practical. A local adversary can recover the ECDSA key from Intel fTPM in 4-20 minutes depending on the access level. We even show that these attacks can be performed remotely on fast networks, by recovering the authentication key of a virtual private network (VPN) server in 5 hours.