CPU, APU & Chipsets News - Page 10

Intel CEO Pat Gelsinger said he hopes to one day make chips for its competitor AMD, and that it has plans on becoming the world's leading foundry by 2030.

During the recent IFS Direct Connect event, Intel CEO Pat Gelsinger said he wants the company to enter the custom chips market -- something AMD is heavily involved in, but has its semi-custom chips made in Taiwan at TSMC -- as well as supply Intel's process nodes and "sharing" each stage of manufacturing technologies, including its exciting new packaging technology.

Intel CEO Pat Gelsinger said: "Well, if you go back to the picture I showed today, Paul, there are Intel products and Intel foundry, There's a clean line between those, and as I said on the last earnings call, we'll have a setup separate legal entity for Intel foundry this year. We'll start posting separate financials associated with that going forward. And the foundry team's objective is simple: Fill. The. Fabs. Deliver to the broadest set of customers on the planet".

Intel's next-gen Xeon "Clearwater Forest" CPUs will feature up to 288 cores based on the new Darkmont CPU architecture, with no P-Cores in sight with Clearwater Forest, its 288 E-Cores.

The latest information on Intel's next-gen Clearwater Forest CPUs is that they will feature Foveros Direct advanced packaging technology, with 3D stacking technology allowing Intel to drive the E-Core count up to an incredible 288 cores of processing power.

Intel's latest hybrid bonding technology is codenamed Foveros Direct for 3D stacking of the Clearwater Forest-based Xeon CPUs. This CPU packaging technology features a base time on top of the interposer, that's connected through a high-speed I/O, EMIB, and the cores will be sitting on the highest layer of the chip.

AMD's CEO Lisa Su has been confirmed as giving the opening keynote at Computex 2024.

The show, which runs from June 4 to June 7 in Taipei, is themed around 'connecting AI' (to no one's surprise) and Lisa Su will be given the opportunity to open the proceedings (as VideoCardz flagged up). That opening keynote will actually happen on June 3.

What will the CEO's speech contain in the way of revelations? The immediate obvious possibility that springs to mind is an announcement of Zen 5, which in its desktop incarnation will be either Ryzen 8000 or 9000 processors.

Intel has penciled in that it will be making the world's best chips later this year, beating TSMC, and teasing its next-gen Intel 14A process node that will go into production in 2026.

The latest Intel 18A process node is gearing up right now, with Microsoft being the first customer with a next-gen processor made on Intel 18A, but the company was confident in announcing its new Intel 14A process node that will cement its position as the fastest chip maker in the world, toppling TSMC, in 2026.

The news was announced at the Intel Foundry Services (IFS) Direct Connect 2024 event. Intel said its current Intel 7 and Intel 4 process nodes are in the market, with its upcoming Intel 3 preparing for high-volume manufacturing (HVM).

Intel held its first-ever Intel Foundry Direct Connect event, with the chipmaker getting customers, ecosystem partners, and leaders across the technology industry to be in person at the event.

Some of the speakers at the Intel Foundry Direct Connect event included Microsoft CEO Satya Nadella, US Secretary of Commerce Gina Raimondo, Arm CEO Rene Haas, OpenAI CEO Sam Altman, and more. At the event, Microsoft CEO Satya Nadella announced that Microsoft has chosen Intel's new 18A process node for a new chip is wants to make.

Microsoft CEO Satya Nadella explained: "We are in the midst of a very exciting platform shift that will fundamentally transform productivity for every individual organization and the entire industry. To achieve this vision, we need a reliable supply of the most advanced, high-performance and high-quality semiconductors. That's why we are so excited to work with Intel Foundry, and why we have chosen a chip design that we plan to produce on Intel 18A process".

AMD's just-released Ryzen 8040 series "Hawk Point" processors will have their included NPUs (Neural Processing Unit) receiving real-time monitoring inside of Windows Task Manager, according to AMD.

In a new post on its website, AMD explained: "AMD has been working with Microsoft to enable MCDM (Microsoft Compute Driver Model) infrastructure on the AMD NPU (Neural Processing Unit)-enabled Ryzen 8040 Series of mobile processors. MCDM is a derivative of Windows Display Driver Model (WDDM) that is targeting non-GPU, compute devices, such as the NPU. MCDM enables NPUs to make use of the existing GPU device management infrastructure, including scheduling, power management, memory management, and performance debugging with tools such as the Task Manager. MCDM serves as a fundamental layer, ensuring the smooth execution of AI workloads on NPU devices".

AMD explains: "One of the reasons AMD has integrated an NPU into select Ryzen 8040 processors is to help vendors create new, AI-powered apps and experiences. Being able to monitor device usage in real-time can make software development easier, so integrating this functionality makes sense for everyone, from developers and system vendors to the individual end-user".

AMD's next-gen Zen 6-based consumer processors are sounding exciting, with new rumors suggesting the codenamed "Medusa" CPUs will feature next-gen RDNA 5 integrated GPUs and 2.5D interconnects, moving away from the traditional multi-die design.

In a new post on X, leaker Everest said that Medusa features an integrated RDNA 5 GPU, which sees AMD skipping over RDNA 4 for Zen 6. Medusa is the codename of the Zen 6 client processors, and the leaker previously teased that Medusa will be using a 2.5D interconnect with "much higher bandwidth".

AMD's next-gen Zen 5 architecture will be unleashed later this year with both the Granite Ridge and Strix Point APU families, with Granite Ridge CPUs featuring integrated RDNA 2-based GPU, as both architectures share the same I/O die where the GPU is located. AMD's next-gen Strix Point APU will feature an upgraded Navi 3.5 (RDNA 3.5) GPU design that we can't wait to see what can do on an APU.

Intel's new Core i9-14900KS processor isn't too far away now, with the limited edition CPU hitting a French retailer early and priced at over $800. Check it out:

In a retail listing by French retailer PC21.FR, the new Intel Core i9-14900KS limited edition processor costs €768.34 (&euro 640.29 without taxes), and the latter costs &euro 752.62 (&euro 627.19 without taxes) in either boxed or tray variant, respectively. This works out to around $827 USD (without taxes) and around $810 with taxes, respectively.

Intel's upcoming Core i9-14900KS processor packs 24 cores and 32 threads of CPU power spread across 8 P-Cores and 16 E-Cores, matching the Core i9-14900K and Core i9-14900KF (KF variant drops the integrated GPU). Intel's upcoming Core i9-14900KS processor packs 24 cores and 32 threads of CPU power spread across 8 P-Cores and 16 E-Cores, matching the Core i9-14900K and Core i9-14900KF (KF variant drops the integrated GPU).

AMD's Zen 5 processors are set to enter mass production in Q3 according to a new report.

This is from UDN, a Chinese tech site, as highlighted by Dan Nystedt as you can see in the tweet above (which was noticed by Wccftech).

UDN tells us that TSMC will start making Zen 5 processors in Q2, but that volume production won't be fully spun up until Q3 of 2024.

Intel's next-gen Lunar Lake CPU has been leaked, showing off an 8-core, 8-thread configuration with cache sizes, and we now have a CPU boost number to scope out. Check it out:

The next-gen Intel Lunar Lake CPUs will be for ultra-low mobile power systems, which will succeed the current-gen Core Ultra 100U "Meteor Lake" CPUs that we've seen deployed since CES 2024 earlier this year. Intel will have its Arrow Lake CPUs hitting the desktop processor market later this year, succeeding the Raptor Lake Refresh CPUs that make up the 14th Gen Core CPU family, while Lunar Lake will enter the low-power laptop market.

The new leak on Intel's upcoming Lunar Lake CPUs sees it in A1 stepping form, with 8 cores and 8 threads (lacking Hyper-Threading), which is something that the Arrow Lake CPU family will also reportedly see (a lack of HT). We see CPU cache sizes here, with 836KB of L1 cache (112KB per P-Core, 96KB per E-Core), 14MB of L2 cache (2.5MB per P-Core, 4MB per four E-Cores), and 12MB of L3 cache in total.