The chip maker was selected for an Energy Department project meant to show American tech independence. But problems at Intel have thrown a wrench into the effort. From a report: When it selected Intel to help build a $500 million supercomputer last year, the Energy Department bet that computer chips made in the United States could help counter a technology challenge from China. Officials at the department's Argonne National Laboratory predicted that the machine, called Aurora and scheduled to be installed at facilities near Chicago in 2021, would be the first U.S. system to reach a technical pinnacle known as exascale computing. Intel pledged to supply three kinds of chips for the system from its factories in Oregon, Arizona and New Mexico. But a technology delay by the Silicon Valley giant has thrown a wrench into that plan, the latest sign of headwinds facing government and industry efforts to reverse America's dependence on foreign-made semiconductors. It was also an indication of the challenges ahead for U.S. hopes to regain a lead in critical semiconductor manufacturing technology. Intel, which supplies electronic brains for most personal computers and web services, has long driven miniaturization advances that make electronic devices smaller, faster and cheaper. But Robert Swan, its chief executive, warned last month that the next production advance would be 12 months late and suggested that some chips for Aurora might be made outside Intel factories. Intel's problems make it close to impossible that Aurora will be installed on schedule, researchers and analysts said. And shifting a key component to foreign factories would undermine company and government hopes of an all-American design. "That is part of the story they were trying to sell," said Jack Dongarra, a computer scientist at the University of Tennessee who tracks supercomputer installations around the world. "Now they stumbled."

Read more of this story at Slashdot.

Last month Intel's stock lost $50 billion in valuation — while the valuation for Taiwan-based TSMC jumped by over 50%. The former chief of staff to Intel CEO Andrew Grove (and later general manager of Intel China) explains why this moment was 15 years in the making: Learning curve theory says that the cost of manufacturing a product declines as the volume increases. Manufacturing chips for the whole computer industry gave Intel huge advantages of scale over any other semiconductor manufacturer and resulted in the company becoming the world's largest chip manufacturer with enviable profit margins. Chaos theory says that a small change in one state of a system can cause a large change in a later stage. In Intel's case, this was not getting selected by Apple for its iPhones. Each successive era of computing was 10x the size of the previous era, so while Intel produced hundreds of millions of microprocessors per year, the mobile phone industry sells billions of units per year. Apple's decision in 2005 to use the ARM architecture instead of Intel's gave Taiwan-based TSMC, the foundry chosen to manufacture the processor chips for the iPhone, the learning curve advantage which over time enabled it to pull ahead of Intel in manufacturing process technology. Intel's integrated model, its competitive advantage for decades, became its vulnerability. TSMC and ARM created a tectonic shift in the semiconductor industry by enabling a large number of "fabless" chip companies such as Apple, AMD, Nvidia and Qualcomm, to name a few. These fabless companies began to out-design Intel in the mobile phone industry and accelerated TSMC's lead over Intel in high volume manufacturing of the most advanced chips. Samsung, which also operates a foundry business, has been another beneficiary of this trend.

Read more of this story at Slashdot.

"I hope AVX512 dies a painful death," Linus Torvalds said last month, "and that Intel starts fixing real problems instead of trying to create magic instructions to then create benchmarks that they can look good on." Friday PC World published some reactions from Intel: Torvalds wasn't the only person to kick AVX-512 in the shins either. Former Intel engineer Francois Piednoel also said the special instruction simply didn't belong in laptops, as the power and die space area trade-offs just aren't worth it. But Intel Chief Architect Raja Koduri says their community loves it because they're seeing a huge performance boost: "AVX-512 is a great feature. Our HPC community, AI community, love it," Koduri said, responding to a question from PCWorld about the AVX-512 kerfuffle during Intel's Architecture Day on August 11. "Our customers on the data center side really, really, really love it." Koduri said Intel has been able to help customers achieve a 285X increase in performance in "our good old CPU socket" just by taking advantage of the extension... Koduri acknowledged some validity to Torvald's heat, too. "Linus' criticism from one angle that 'hey, are there client applications that leverage this vector bit yet?' may be valid," he said. Koduri explained further that Intel has to maintain a hardware software contract all the way from servers to laptops, because that's been the magic of the ecosystem. "(That's) the great thing about the x86 ecosystem, you could write a piece of software for your notebook and it could also run on the cloud," Kodori said. "That's been the power of the x86 ecosystem..." And no, hate on AVX-512 and special instructions all you want, Intel isn't going to change direction. Koduri said it will continue to lean on AVX-512 as well as other instructions. "We understand Linus' concerns, we understand some of the issues with first generation AVX-512 that had impact on the frequencies etc, etc," he said "and we are making it much much better with every generation." They also summarize some performance testing by blogger Travis Downs, saying it found AVX-512 "doesn't appear to enforce much of a penalty at all on a laptops. Downs' testing found the clock speed only dropped 100MHz when using one active core under AVX-512. "At least, it means we need to adjust our mental model of the frequency related cost of AVX-512 instructions," Downs concluded. "Rather than 'generally causing significant downclocking,' on this Ice Lake chip we can say that AVX-512 causes insignificant or zero licence-based downclocking and I expect this to be true on other Ice Lake client chips as well."

Read more of this story at Slashdot.

Intel on Thursday disclosed a new method for making transistors on semiconductors that its chief architect said could boost the performance Intel's next round of processors by as much as 20%. From a report: The Santa Clara, California-based company is one of the few remaining in the world that both designs and manufactures its own chips. But its manufacturing operations have become a concern among investors after Intel last month said that its next-generation chip-making process, called its 7-nanometer process node, would be delayed. Analysts believe the delays could cement the lead that rivals such as Taiwan Semiconductor Manufacturing Co have gained in making smaller, more power efficient chips. Intel's shares have fallen nearly 20% since the delays were disclosed. On Thursday, Intel sought to buck the notion that the single-number names given to each generation of chip process node tell the entire story by disclosing improvements on its existing 10-nanonmeter process node. It announced a new way of making what it now calls "SuperFin" transistors, which, along with a new material being used to improve the capacitors on chips, is expected to boost the performance of Intel's forthcoming processors, despite their still being made on 10-nanometer manufacturing lines.

Read more of this story at Slashdot.