CPU, APU & Chipsets News - Page 10
Intel Core Ultra is finally here, and it powers over 230 of the 'world's first AI PCs'
Using the new Meteor Lake technology, Intel's latest Core Ultra processors present a major architectural shake-up for the company - one focused (at the moment) on the thin and powerful laptop market. The first generation of Core Ultra CPUs that use the new and more efficient 7nm Intel 4 process are finally here, powering the first wave of AI PCs from companies like ASUS, MSI, Lenovo, Dell, GIGABYTE, and more.
What makes the new Intel Core Ultra range tailor-made for AI PCs compared to what has come before? That comes down to the major architectural shark-up - the biggest in Intel's history. Meteor Lake, which is at the heart of this first wave of Core Ultra products, combined multiple tiles (or chips) into a single chip - covering an AI-powered SoC Tile, an Intel Arc Graphics GPU Tile, and a dedicated I/O Tile for Thunderbolt and high-speed PCIe Gen5.
"In the next four years, AI PCs will comprise 80% of the PC market, and together with our vast ecosystem of hardware and software partners, Intel is best positioned to deliver this next generation of computing," Michelle Johnston Holthaus, Intel executive vice president and general manager of Client Computing Group said.
TSMC is developing 1.4nm chips called A14 that are due to arrive around 2027
As the size of nodes for chips decreases and their complexity increases, only some companies with the technology can continue to push and produce the most cutting-edge microprocessors. With that said, one stands out from the pack - TSMC.
Earlier today (or was that last night), we reported on the company's upcoming 2nm (or N2) process node that is rumored to power the next-generation iPhone 17 Pro. It's on track for 2025. Naturally, the company's portfolio extends beyond this, and even though it's assumed that both TSMC and its competitors like Samsung that the development of 1.4nm process nodes will follow 2nm - we've now for confirmation of its existence via a company slide showcased during a recent Future of Logic panel.
As posted by Dylan Patel over on X, TSMC's 1.4nm node or A14 is currently "in development" alongside the company acknowledging that AI will power microprocessors for the foreseeable future. With 2nm expected in 2025, which, like 3nm and the iPhone 15, will probably debut alongside a new round of Apple products.
TSMC shows off 2nm chips to Apple, mass production expected in 2025 for the iPhone 17 Pro
Apple announced its new iPhone 15 Pro smartphone earlier this year, powered by its new A17 Pro chip that's made on TSMC's 3nm process, but its future iPhone 17 Pro will reportedly be fabbed on TSMC's new 2nm process node.
The news is coming from a new report, TSMC will begin mass production of its new 2nm node in 2025, where Apple will be the first client, and its next-gen iPhone 17 Pro smartphone will be the first with a 2nm chip. Apple is TSMC's largest client, taking in Taiwan Semiconductor Manufacturing Company's entire supply of 3nm chips in 2023, where TSMC gave Apple first dibs on its new node before its competitors.
In a new report from the Financial Times, they explained: "TSMC, which dominates the global market in processors, has already shown the process test results for its "N2" - or 2 nanometre - prototypes to some of its biggest customers, including Apple and NVIDIA, according to two people with direct knowledge of the discussions".
Intel CEO will deliver CES 2024 keynote on January 9, teases 'AI everywhere starts with Intel'
Intel has announced that it will have its CEO Pat Gelsinger on stage at CES 2024, providing a keynote, where CNBC correspondent Kristina Partsinevelos will join to talk about all things AI.
We know that AI will be a focus of all major tech companies moving forward, with Intel's upcoming "AI Everywhere" event penciled in for December 14, which is just a few days away. We will find out more details about Intel's new Core 100 series codenamed Meteor Lake, which is one of the more interesting CPU designs the company has pumped out, and all eyes are focused on it.
AI, on the other hand, is the core focus of Intel right now... however, the CES 2024 keynote will see Intel talk about its Core 100 series and the Meteor Lake architecture in general and how it is powering waves of new laptops and other Meteor Lake-powered products. Intel will be hosting an Open House at CES 2024 that will be led by Michelle Johnston Holthaus, the GM of Intel CCG (Client Computing Group) where people on the ground will be able to get their hands on some Intel Core Ultra demo systems.
AMD teases its next-gen Strix Point APUs with 3x generative AI NPU performance with XDNA 2
AMD just announced its new Ryzen 8040 series APUs codenamed "Hawk Point," which have improvements in generative AI workloads, with Ryzen AI on board and helping out through a next-gen NPU (Neural Processing Unit). But after this, we've got the next-gen "Strix Point," which will have over 3x the generative AI NPU performance.
AMD itself is promising that its next-gen Ryzen 9000 series APUs codenamed "Strix Point" will have "more than 3x generative AI NPU performance" compared to Strix Point with its new XDNA 2 AI architecture.
Right now, the Ryzen AI cores inside of the Ryzen 7040 "Phoenix" APU series have around 10 TOPS of compute performance for generative AI, compared to the upcoming Ryzen 8040 "Hawk Point" APU series that has 16 TOPS of compute performance for generative AI.
TSMC pumping out Apple 3nm orders: AMD, Intel, NVIDIA, Qualcomm, all jump on TSMC's 3nm in 2024
TSMC is too busy pumping out 3nm chips for Apple's new M3 series processors, with the Taiwanese giant to see capacity utilization decline at the beginning of 2024, but in the second half of 2024, things will pick up as other 3nm customers will join the party. This includes AMD, Intel, NVIDIA, Qualcomm, and MediaTek.
According to Ctee, we are to expect TSMC's current 3nm capacity utilization rate to drop to 65% before things pick up in 2H 2024. US foreign investors are mostly optimistic about TSMC and its generative AI, semiconductor recovery cycle, 2nm technology leadership, and sped-up overseas factory expansion, so they are "optimistic about the long-term investment value" in the company.
Intel will be using TSMC's new N3B process for its next-gen, low-power architecture in Lunar Lake MX (LNL), while we'll also see Intel's upcoming Arrow Lake-H and Arrow Lake-HX processors made on 3nm, which will be "used to significantly fill TSMC's production capacity".
AMD announces Ryzen 8040 series 'Hawk Point' APUs: Zen 4 CPU, RDNA 3 GPU, XDNA for AI
AMD has just detailed its new Ryzen 8040 series APUs that are codenamed Hawk Point, where they will feature Zen 4 CPU cores, RDNA 3 GPU cores, and an NPU (Neural Processing Unit) and its Ryzen AI core.
The new flagship AMD Ryzen 9 8945HS processor features 8 cores and 16 threads, boosting up to 5.2GHz and packing a Radeon 780M RDNA 3-based GPU that features 12 Compute Units and a 35-54W TDP. There are a few different families of Hawk Point APUs here, with 1 x Ryzen 9 and 3 x Ryzen 7 APUs that are all 8-core, 16-thread parts.
Under that, we've got 5 x Ryzen 5 APUs that are 6-core, 12-thread chips, while there's a Ryzen 3 APU with 4 cores and 8 threads at the bottom of the Ryzen 8040 series APU stack. Each of them has on-board graphics and an NPU... except for the last two Ryzen 8040 series CPUs (the Ryzen 5 8540U and Ryzen 3 8440U both don't have NPUs).
Intel's new marketing campaign is aggressive against AMD, calls Ryzen CPUs 'snake oil'
Intel is about to launch its next-gen Meteor Lake platform and has just launched its refreshed 14th Gen Core "Raptor Lake Refresh" CPUs, and now has an aggressive new marketing campaign: comparing AMD Ryzen 5 processors to snake oil.
In that aggressive new marketing campaign is something Intel is calling "Core Truths"; it acts like an information brochure that consumers can use to learn about the latest CPU technologies. Intel specifically calls out specific AMD Ryzen CPUs that are using older CPU architectures... while acting like they don't, and still don't, do that (spoiler alert: oh yes, yes, they do).
AMD and Intel were on good terms when it came to a direct attack-style marketing campaign, similar to how Intel also ran that marketing campaign that said AMD's new EPYC CPU dies were "glued together." Well, Intel was scared of AMD's impending dominance -- in which it has, oh-so-beautifully for years, eaten at Intel's dominance in server and HPC businesses -- where EPYC has kicked Xeon's ass.
AMD Ryzen 7 5700X3D launching in Q1 2024, fantastic gaming CPU for AM4 socket
We've already heard reports of AMD working on X3D processors for its previous-gen AM4 platform, with new sources corroborating the leaks: AMD will be launching its Ryzen 7 5700X3D in the first few months of 2024.
AMD's upcoming Ryzen 7 5700X3D processor will be an 8-core, 16-thread CPU with 96MB of X3D cache on-board, while we're expecting a base clock of 3.0GHz and a boost clock of up to 4.1GHz. This means it is 400MHz slower in frequency over the Ryzen 7 5800X3D, which AMD is probably using slightly gimped 5800X3D chips to make its upcoming 5700X3D.
There's no new information on the Ryzen 5 5500X3D, which is expected to be a more budget X3D offering as a 6-core, 12-thread offering that should have a 3.0GHz base clock and boost clock of up to 4.0GHz. ECSM is confirming this on Twitter, posting up a bunch of SKUs that it sees AMD launching in Q1 2024: Zen 3 GT, Zen 3 X3D, and Ryzen 8000 APUs that we've been hearing about for a while now.
AMD's next-gen EPYC-E CPU: up to 64 x Venice Zen 6 cores, PCIe 6.0 support
AMD's next-gen Zen 6 architecture won't be hitting desktops or servers for a number of years yet... we've still got Zen 5 to get to, folks... but we're hearing about the next-gen EPYC-E processors that will be powered by the next-gen Zen 6 architecture.
The new rumors on AMD's next-gen Zen 6 architecture and EPYC-E processors are coming from leaker Moore's Law is Dead after some details on the architectural side of things for both Zen 5 and Zen 6 a few months ago in September, from a leaked internal AMD slide deck.
The codename for Zen 5 is "Nirvana," while Zen 6 is codenamed "Morpheus" where we're learning that Zen 6 will be increasing core counts up to 32 cores per chiplet (CCD) and supporting high-prevision floating point (FP16) operations acceleration for both AI and Machine Learning (ML).