As wafer costs increase, so are the costs of GPUs. According to a post on the Board Forums, AMD says it's increasing the price of its RX 6000 series GPUs by 10 percent across the board. ExtremeTech reports: This pricing change will apparently occur in the next shipment of GPUs to its partners, which will apparently drive up the price of these GPUs by $20 to $40 USD. This news arrives just in time for the holiday shopping season, when demand for GPUs is expected to increase even more, as if that is even possible.

According to a translation of the board posting, AMD is citing TSMC wafer costs as the reason for the change, and as we reported earlier, sub-16nm prices, including 12nm, 7nm, and 5nm, are said to have increased roughly 10 percent, while TSMC's older nodes have gone up by as much as 20 percent. AMD seems to be passing this price increase along to its partners, who in turn are passing it along to us, the customer, or the scalper, as it were. Then the scalper passes it along to us, the gamers. Although, as Videocardz points out AMD also produces its CPUs at TSMC and there hasn't been a similar across-the-board increase, which is curious.

Years after its failed Steam Machines, Valve is slowly but surely improving the state of Linux gaming. From a report: The company's upcoming Steam Deck handheld runs atop Linux, and its Proton compatibility layer lets it -- and other computers -- play Windows games as well. Now, Valve has officially added support for Nvidia's DLSS machine learning temporal upscaling technique to Proton, potentially bringing big FPS boosts and less flicker in games that support the technology.

Proton 6.3-8 is the first stable release to include support for DLSS, after the feature previously hit experimental builds in October, though it appears you'll still need to set PROTON_ENABLE_NVAPI=1 and dxgi.nvapiHack = False to turn it on. DLSS won't come to the AMD-powered Steam Deck, of course, since it requires proprietary Nvidia machine learning silicon, but we recently learned the Steam Deck will support AMDâ(TM)s arguably much less capable FSR.

An anonymous reader quotes a report from XDA Developers: Last week, we reported that MediaTek is planning to build a chipset for Windows on ARM. As it turns out, the Windows on ARM chipset space could be even hotter than that, because there's a reason that we've only seen Qualcomm SoCs in ARM PCs so far. Qualcomm actually has an exclusivity deal with Microsoft for Windows on ARM, and speaking with people familiar with it, we've learned that the deal is set to expire soon. Other than the fact that Microsoft has publicly said that anyone who wants to can build a Windows on ARM chip, this really shouldn't come as a surprise. Qualcomm didn't just start building PC chips hoping that Microsoft would compile Windows to support it. No, these two companies worked together to make it happen. Because of that, Qualcomm gets to enjoy a bit of exclusivity.

One thing I wasn't able to learn is when the deal will expire, only that it's the thing holding back other chip vendors from competing in the space. It's possible that Samsung might want to throw its hat into the ring with its Exynos processors too, especially given its recent partnership with AMD for graphics power. This is also presumably why Apple Silicon Macs aren't officially supported for running Windows 11, so hopefully that will change as well. [...] Between MediaTek's Executive Summit and Qualcomm's Investor Day, there's been a very clear message that ARM SoC vendors absolutely believe that the 'Wintel' partnership is going to fade and that the transition to ARM isn't just happening, it's inevitable. Naturally, that means that all of these companies are going to want to be part of it when it opens up. Qualcomm has quite a head start though, given that it's been doing this for a few years and on top of that, it's going to start building its own custom silicon thanks to its Nuvia acquisition.

America's push to manufacturer more products domestically gets an in-depth look from CNET — including a new Intel chip factory outside of Phoenix.

CNET calls it a fork in the road "after squandering its lead because of a half decade of problems modernizing its manufacturing..." With "a decade of bad decisions, this doesn't get fixed overnight," says Pat Gelsinger, Intel's new chief executive, in an interview. "But the bottom is behind us and the slope is starting to feel increasingly strong...." More fabs are on the way, too. In an enormous empty patch of dirt at its existing Arizona site, Intel has just begun building fabs 52 and 62 at a total cost of $20 billion, set to make Intel's most advanced chips, starting in 2024. Later this year, it hopes to announce the U.S. location for its third major manufacturing complex, a 1,000-acre site costing about $100 billion. The spending commitment makes this year's $3.5 billion upgrade to its New Mexico fab look cheap. The goal is to restore the U.S. share of chip manufacturing, which has slid from 37% in 1990 to 12% today. "Over the decade in front of us, we should be striving to bring the U.S. to 30% of worldwide semiconductor manufacturing," Gelsinger says...

But returning Intel to its glory days — and anchoring a resurgent U.S. electronics business in the process — is much easier said than done. Making chips profitably means running fabs at maximum capacity to pay off the gargantuan investments required to stay at the leading edge. A company that can't keep pace gets squeezed out, like IBM in 2014 or Global Foundries in 2018. To catch up after its delays, Intel now plans to upgrade its manufacturing five times in the next four years, a breakneck pace by industry standards. "This new roadmap that they announced is really aggressive," says Linley Group analyst Linley Gwennap. "I don't have any idea how they are going to accomplish all of that...."

Gelsinger has a tech-first recovery plan. He's pledged to accelerate manufacturing upgrades to match the technology of TSMC and Samsung by 2024 and surpass them in 2025. He's opening Intel's fabs to other companies that need chips built through its new Intel Foundry Services (IFS). And he's relying on other foundries, including TSMC, for about a quarter of Intel's near-term chipmaking needs to keep its chips more competitive during the upgrades. This three-pronged strategy is called IDM (integrated design and manufacturing) 2.0. That's a new take on Intel's philosophy of both designing and making chips. It's more ambitious than the future some had expected, in which Intel would sell its factories and join the ranks of "fabless" chip designers like Nvidia, AMD and Qualcomm that rely on others for manufacturing...

Shareholders may not like Gelsinger's spending-heavy strategy, but one community really does: Intel's engineers... Gelsigner told the board that Intel is done with stock buybacks, a financial move in which a company uses its cash to buy stock and thereby increase its price. "We're investing in factories," he told me. "That's going to be the use of our cash...."

"We cannot recall the last time Intel put so many stakes in the ground," said BMO Capital Markets analyst Ambrish Srivastava in a July research report after Intel announced its schedule.
Intel will even outpace Moore's law, Gelsinger tells CNET — more than doubling the transistor count on processors every two years. "I believe that you're going to see from 2025 to 2035 a very healthy period for Moore's Law-like behavior."

Although that still brings some risk to Intel's investments if they have to pass the costs on to customer, a Linley Group analyst points out to CNET. "Moore's Law is not going to end when we can't build smaller transistors. It's going to end when somebody says I don't want to pay for smaller transistors."

According to a report from Nikkei Asia (paywalled), "The Japanese-made Fugaku captured its fourth consecutive title as the world's fastest supercomputer on Tuesday, although a rival from the U.S. or China is poised to steal the crown as soon as next year." From a report: But while Fugaku is the world's most powerful public supercomputer, at 442 petaflops, China is believed to secretly operate two exascale (1,000 petaflops) supercomputers, which were launched earlier this year. The top 10 list did not change much since the last report six months ago, with only one new addition -- a Microsoft Azure system called Voyager-EUS2. Voyager, featuring AMD Epyc CPUs and Nvidia A100 GPUs, achieved 30.05 petaflops, making it the tenth most powerful supercomputer in the world.

The other systems remained in the same position - after Japan's Arm-based Fugaku comes the US Summit system, an IBM Power and Nvidia GPU supercomputer capable of 148 petaflops. The similarly-architected 94 petaflops US Sierra system is next. Then comes what is officially China's most powerful supercomputer, the 93 petaflops Sunway TaihuLight, which features Sunway chips. The Biden administration sanctioned the company earlier this year. You can read a summary of the systems in the Top10 here.

Microsoft has announced a new edition of Windows 11 designed specifically for the K-8 education sector, dubbed "Windows 11 SE." This new edition of Windows 11 is designed to address fundamental challenges that schools are facing day to day with improved performance, optimized resources, and simple to deploy and manage. From a report: Microsoft says Windows 11 SE has been optimized for education focused low-cost PCs, most of which start at the affordable price of $249 and are powered by low-end Intel and AMD chips. Windows 11 SE was designed with feedback from teachers and school IT admins in mind. Unlike normal Windows 11, Windows 11 SE comes pre-loaded with Microsoft Office out of the box, including Word, PowerPoint, Excel, OneNote, and OneDrive, which can also be used offline as part of a Microsoft 365 license.

Microsoft has also limited some of the multitasking features, including reducing the amount of apps that can be snapped on screen at once to just two; side by side. The Microsoft Store app is also disabled. Windows 11 SE also automatically runs apps in full-screen, which makes sense considering most Windows 11 SE PCs will feature small 11-inch displays. It also removes access to the "This PC" area in File Explorer by default, as it's an area most students don't need to access when working on school work. Windows 11 SE is "cloud backed" meaning it will mirror all your saved documents stored locally to the cloud.

Advanced Micro Devices said Meta Platforms, formerly known as Facebook, is becoming a user of its server processors, further eroding Intel's hold on that lucrative market. From a report: Meta will use AMD Epyc processors in its data center computers, the two companies said Monday at an event. AMD also unveiled a new version of that chip with extra memory, which Microsoft Corp. will use in an offering from its Azure cloud computing service. The chipmaker also showed off a new graphics chip for artificial intelligence workloads and gave hints about its next generation of processors coming in 2022. The addition of Meta, the world's largest social media company, to AMD's customer list means it now supplies all the top operators of the giant computing networks that run the internet. Winning those major spenders was part of Chief Executive Officer Lisa Su's plan to resurrect AMD and have it reach market share levels it had only briefly flirted with amid years of struggling to keep up with Intel.

"Intel is gearing up to go to a war with Nvidia," writes Slashdot reader labloke11. "They have their OneAPI and their GPU. It will be interesting... For me, I like competition." Phoronix reports: While Intel Alder Lake is dominating today's news cycle, Intel and Microsoft also announced today that they have brought oneAPI Level Zero and Intel OpenCL support to Windows Subsystem for Linux (WSL2) while employing Intel graphics hardware acceleration. Similar to NVIDIA bringing CUDA and their accelerated GPU support to WSL2 as well as similar efforts by AMD on the Radeon side, Intel and Microsoft are now having Intel graphics compute working within the Linux confines on Windows 11 or Windows 10 21'H2. Hardware-accelerated oneAPI Level Zero, OpenVINO, and OpenCL on Intel graphics hardware can now be enjoyed within the WSL2 environment when using the latest updates and drivers. Like with the rest of the WSL2 stack and capabilities from other GPU vendors, this is at a near-native level of performance. More information can be found via the Microsoft Command Line blog and Intel blog.

MojoKid writes: After months of speculation and early look teases, Intel's 12th Gen Core processors are finally ready for prime time. Today marks the embargo lift for independent reviews of Alder Lake and it's clear Chipzilla is back and bringing the fight again versus chief rival AMD. Intel 12th Gen Core processors incorporate two new CPU core designs, dubbed Efficiency (E-core) and Performance (P-core). In addition to this new hybrid core architecture, 12th Gen Core processors and the Z690 motherboard chipset platform also feature support for the latest memory and IO technologies, including PCI Express Gen 5, DDR5, Thunderbolt 4 and Wi-Fi 6E. The new Core i9-12900K features a monolithic, 16-core (24-thread) die with 8 Performance cores and 8 Efficiency cores, while the Core i5-12600K has 10 cores/16-threads, comprised of 6 P-cores and 4 E-cores. Alder Lake E-cores don't support HyperThreading but P-cores can process two threads simultaneously while E-cores can manage only one, hence the asymmetric core counts.

In the benchmarks, the 16-core Core i9-12900K doesn't sweep AMD's 16-core Ryzen 9 5950X across the board in multi-threaded tests, but it certainly competes very well and notches plenty of victories. In the lightly-threaded tests though, it's a much clearer win for Intel and gaming is an obvious strong point as well. Alder Lake's performance cores are as fast as they come. The $589 (MSRP) 16-core Core i9-12900K competes well with the $750 16-core Ryzen 9 5950X, and the $289 10-core Core i5-12600K has a lower MSRP than a $299 6-core Ryzen 5 5600X. The new Core i5's power and performance look great too, especially when you consider this $289 chip outruns Intel's previous-gen flagship Core i9-11900K more often than not, and it smokes a Ryzen 5 5600X.

NotebookCheck: It seems that Intel was not telling the full story when it compared the Core i9-12900K against the Ryzen 9 5950X, currently AMD's most comparable processor. As it turns out, Intel allowed the Core i9-12900K to consume 2.4x the power of its AMD competitor and benchmarked the Ryzen 9 5950X using an older version of Windows 11 with AMD performance issues.

Phoronix: Now that Windows 11 has been out as stable and the initial round of updates coming out, I've been running fresh Windows 11 vs. Linux benchmarks for seeing how Microsoft's latest operating system release compares to the fresh batch of Linux distributions. First up is the fresh look at the Windows 11 vs. Linux performance on an Intel Core i9 11900K Rocket Lake system. Microsoft Windows 11 Pro with all stable updates as of 18 October was used for this round of benchmarking on Intel Rocket Lake. The Windows 11 performance was being compared to all of the latest prominent Linux distributions, including: Ubuntu 20.04.3 LTS, Ubuntu 21.10, Arch Linux (latest rolling), Fedora Workstation 35, Clear Linux 35150. All the testing was done on the same Intel Core i9 11900K test system at stock speeds (any frequency differences reported in the system table come down to how the information is exposed by the OS, i.e. base or turbo reporting) with 2 x 16GB DDR4-3200 memory, 2TB Corsair Force MP600 NVMe solid-state drive, and an AMD Radeon VII graphics card.

Each operating system was cleanly installed and then run at its OS default settings for seeing how the out-of-the-box OS performance compares for these five Linux distributions to Microsoft Windows 11 Pro. But for the TLDR version... Out of 44 tests run across all six operating systems, Windows 11 had just three wins on this Core i9 11900K system. Meanwhile Intel's own Clear Linux platform easily dominated with coming in first place 75% of the time followed by Fedora Workstation 35 in second place with first place finishes 9% of the time. The geometric mean for all 44 tests showed Linux clearly in front of Windows 11 for this current-generation Intel platform. Ubuntu / Arch / Fedora were about 11% faster overall than Windows 11 Pro on this system. Meanwhile, Clear Linux was about 18% faster than Windows 11 and enjoyed about 5% better performance overall than the other Linux distributions.

An anonymous reader quotes a report from ZDNet: The reign of Apple's M1 SoC at the top of the Geekbench speed benchmarks may soon be over with the impending arrival of Intel's 12th-generation Alder Lake mobile processors. Hardware site Wccftech appears to have been leaked Intel's upcoming Core i9-12900HK mobile CPU, and has now revealed the first benchmarks. The results show Intel's mobile CPU narrowly outperforming Apple's flagship 10-core M1 Max, which also integrates a 32-core GPU and 64GB of unified memory.

In these latest tests, the Core i9-12900HK outperforms the M1 Max on both single-core and multi-core benchmarks. The margin is slim, but is important for Intel since Apple ditched its CPUs for its own designs in new MacBooks. Intel's Alder Lake CPU didn't beat the M1 Max by much, with respective single-core scores of 1851 and 1785. It beat the Core i9-11980HK and AMD's top mobile CPU, the Ryzen 5980HX, by a bigger margin: the latter two CPUs saw scores of 1616 and 1506, respectively. In the multi-core benchmark, the Core i9-12900HK scored 13256 versus the M1 Max's score of 12753. Again, it trounced AMD's 5980HX, which scored 8217. Wccftech's Alder Lake benchmarks were run using Windows 11, so it's possible Thread Director's hardware scheduling influenced the results.

AMD and Microsoft have issued patches to address the slowdowns reported with Ryzen processors when Windows 11 launched. Engadget reports: The latest chipset driver (version 3.10.08.506) should take care of the UEFI CPPC2 issue, which in some cases didn't "preferentially schedule threads on a processor's fastest core," AMD said. That could have slowed down apps that are sensitive to CPU thread performance. AMD noted that the problem was likely more noticeable in more powerful processors with more than eight cores and 65W or higher Thermal Design Power (TDP).

Meanwhile, Microsoft is rolling out a software update tackling a bug that increased L3 cache latency. The issue impacted apps that need quick memory access, which in turn caused CPUs to slow down by up to 15 percent. The patch, Windows 11 update KB5006746, will be available starting today, but at the time of writing, a page containing instructions for installing it isn't yet live. You should be able to install it via Windows Update too.

The original Xbox was released 20 years ago next month, and to mark the upcoming anniversary, the console's designer has apologized to AMD's engineers and its CEO for Microsoft's last-minute decision to drop AMD for rival Intel. GameSpot reports: Seamus Blackley apologized on Twitter to the AMD engineers who worked with Microsoft to create the prototype Xbox consoles that the company used in the lead-up to the OG Xbox's release in November 2001. To AMD CEO Lisa Su, Blackley said, "I beg mercy." "I was standing there on the stage for the announcement, with [Bill Gates], and there they were right there, front row, looking so sad," he said of AMD engineers in the room. "I'll never forget it. They had helped so much with the prototypes. Prototypes that were literally running the launch announcement demos ON AMD HARDWARE." "I felt like such an ass," Blackley said. Microsoft dropped AMD in favor of Intel due to "pure politics," Blackley said in another tweet.

AMD warned last week that its chips are experiencing performance issues in Windows 11, and now Microsoft's first update to its new OS has reportedly made the problems worse. From a report: TechPowerUp reports that it's seeing much higher latency, which means worse performance, after the Windows 11 update went live yesterday. AMD and Microsoft found two issues with Windows 11 on Ryzen processors. Windows 11 can cause L3 cache latency to triple, slowing performance by up to 15 percent in certain games. The second issue affects AMD's preferred core technology, that shifts threads over to the fastest core on a processor. AMD says this second bug could impact performance on CPU-reliant tasks. TechPowerUp measured the L3 cache latency on its Ryzen 7 2700X at around 10ns, and Windows 11 increased this to 17ns. "This was made much worse with the October 12 'Patch Tuesday' update, driving up the latency to 31.9ns," says TechPowerUp. That's a huge jump, and the exact type of issue AMD warned about.

AMD said Windows 11 can cut game performance on Ryzen CPUs by 10 to 15 percent, and the operating system may not utilize AMD's "preferred core" technology, but a fix is in the works. PCWorld reports: A support note on AMD's web site published this week said Windows 11 may increase L3 cache latency a whopping threefold, which can cause slowdowns in latency sensitive applications. Lighter duty, cache-sensitive games might see a 3 percent to 5 percent hit, and lighter-duty games as e-sports titles could see frame rates drop from 10 to 15 percent. AMD also said its "preferred core" feature, which tells the operating system which core in each CPU can hit the highest clock, also doesn't work right in Windows 11. Each CPU is tested to see which core will run the fastest at the factory and is marked so the OS will dispatch tasks to that "preferred core." Since Windows 11 doesn't seem to work with it right now, any performance bump from using the best core wouldn't happen. The company said the performance cost would be most noticeable in CPUs with more than 8 cores and with TDP ratings above 65 watts.

An anonymous reader quotes a report from The Register: The European Processor Initiative (EPI) has run the successful first test of its RISC-V-based European Processor Accelerator (EPAC), touting it as the initial step towards homegrown supercomputing hardware. EPI, launched back in 2018, aims to increase the independence of Europe's supercomputing industry from foreign technology companies. At its heart is the adoption of the free and open-source RISC-V instruction set architecture for the development and production of high-performance chips within Europe's borders. The project's latest milestone is the delivery of 143 samples of EPAC chips, accelerators designed for high-performance computing applications and built around the free and open-source RISC-V instruction set architecture. Designed to prove the processor's design, the 22nm test chips -- fabbed at GlobalFoundries, the not-terribly-European semiconductor manufacturer spun out of AMD back in 2009 -- have passed initial testing, running a bare-metal "hello, world" program as proof of life.

It's a rapid turnaround. The EPAC design was proven on FPGA in March and the project announced silicon tape-out for the test chips in June -- hitting a 26.97mm2 area with 14 million placeable instances, equivalent to 93 million gates, including 991 memory instances. While the FPGA variant, which implemented a subset of the functions of the full EPAC design, was shown booting a Linux operating system, the physical test chips have so far only been tested with basic bare-metal workloads -- leaving plenty of work to be done. Earlier today, the UK government released its 10-year plan to make the country a global "artificial intelligence superpower," seeking to rival the likes of the U.S. and China. "The so-called 'National Artificial Intelligence Strategy' is designed to boost the use of AI among the nation's businesses, attract international investment into British AI companies and develop the next generation of homegrown tech talent," reports CNBC.

AMD's CFO Devinder Kumar has commented that AMD stands ready to manufacture Arm chips if needed, noting that the company's customers want to work with AMD on Arm-based solutions. From a report: Kumar's remarks came during last week's Deutsche Bank Technology Conference, building on comments from AMD CEO Lisa Su earlier in the year that underscored the company's willingness to create custom silicon solutions for its customers, be they based on x86 or Arm architectures. Intel also intends to produce Arm and RISC-V chips, too, meaning that the rise of non-x86 architectures will be partially fueled by the stewards of the dominant x86 ecosystem. "But I'll tell you from my standpoint, when you look at compute solutions, whether it's x86 or ARM or even other areas, that is an area for our focus on investment for us," AMD CFO Devinder Kumar responded to a question about the company's view of competing Arm chips. "We know compute really well. Even ARM, as you referenced, we have a very good relationship with ARM. And we understand that our customers want to work with us with that particular product to deliver the solutions. We stand ready to go ahead and do that even though it's not x86, although we believe x86 is a dominant strength in that area."

An anonymous reader quotes a report from PCWorld: The latest version of Radeon Software adds an unusual (and welcome) new twist: The ability to automatically overclock your Ryzen processor if you're rocking an all-AMD gaming desktop. Yes, your GPU software can speed up your CPU now, too -- and it can do it all with a single click. [...] The addition of Ryzen auto-overclocking in Radeon Software 21.9.1 continues the theme, and might just allow you to ditch AMD's separate Ryzen Master tool if you're running a Team Red graphics card. They'll need to be newer hardware, though, as the feature currently only supports AMD's latest Ryzen 5000 CPUs and Radeon RX 6000 GPUs.

AMD's blog describes how to use the new tool: "To access this easy-to-use feature, open up Radeon Software using the hotkey 'ALT' + 'R', navigate to the 'Performance' tab found at the top of the window, and select 'Tuning' in the sub tab directly below it. If you have the latest generation of AMD Ryzen and Radeon product installed on your system, a 'Tuning Control' section should appear for your system, allowing you to select 'Auto Overclock' to increase performance on both your processor and graphics card. We also have a new tuning section for CPUs, allowing you to overclock just your CPU. When the feature is selected, the system will ask for a restart and once you are back in Windows, you will be good to go!" "Radeon Software 21.9.1 also adds official Windows 11 support and the ability for Radeon RX 5000-series GPUs to tap into Smart Access Memory," adds PCWorld. "AMD also took the time to tout FidelityFX Super Resolution's rapid uptake. The DLSS rival is now supported in 27 games, with Arkane's awesome-looking Deathloop set to launch this week with native FSR support in place."

You can download these new drivers here.

Intel has pivoted on its server strategy in order to fight a supply-constrained AMD, reports DigiTimes. It's reportedly flooding the market with chips at discount pricing, rather than sticking to MSRP. From a report: While some reports point toward a relative normalization on AMD's CPU supply, AMD has two distinct disadvantages when compared to Intel: It has fewer revenue sources than its much bigger CPU rival, and AMD doesn't own the factories that produce its market-turning Zen chips. Intel, on the other hand, can leverage its vertical integration (meaning that development and manufacturing takes place in an almost entirely Intel-owned and managed supply chain), as well as its massive revenue advantage, to play with final client pricing. In other words, Intel pull a lot more levers to increase demand and (Intel hopes) attract would-be AMD clients back into the Intel fold.

AMD has seemingly been making strides in server market penetration. As seen in renowned system distributor Puget Systems' statistics, AMD has risen from a 5% share in systems sold since June 2020, up to a dominating 60% as of June 2021. However, unserved demand means that companies looking to invest in their server infrastructure or who aim to deploy AMD chips in any major way sometimes can't wait for the chips to become available. And Intel is smartly making it more attractive for those companies to go back to the Intel fold, or to skip AMD in the first place.