CPU, APU & Chipsets News - Page 13
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).
Intel's next-gen Clearwater Forest Xeon CPU has only E-Cores, but up to 288 cores
Intel's next-gen Xeon, codenamed Clearwater Forest, will feature up to 288 cores with an updated CPU architecture, but those 288 cores will exclusively be E-Cores, not the higher-performance P-Cores that Intel has on its CPUs.
This is according to new leaks on the scene, which tease that Intel's 2nd Gen Xeon CPU family will be based on the E-Core CPU architecture, while the first E-Core-only CPUs are codenamed Sierra Forest, which Intel will release in the 2H of 2024. If Intel stays on track, that means we should expect to see Clearwater Forest CPUs sometime in 2025... but this could slip into 2026 easily.
Intel's new Clearwater Forest architecture will feature an updated version of the Skymont E-Core architecture, which it'll be the 5th Gen E-Core architecture, with Gracemont being the first. Intel's next-gen Gracemont architecture will be launching on the new Meteor Lake CPUs, which will be making their debut on December 14 at Intel's upcoming "AI Everywhere" event.
AWS Graviton4 announced: the most powerful and energy-efficient chip Amazon has ever built
AWS has just announced its new Graviton4 processor, the fourth-generation Arm-based server CPU that has 50% more cores, 75% more memory bandwidth, and an overall compute performance uplift of up to 30%.
Amazon Web Services (AWS) was previously manufacturing its Graviton3 and Graviton3E processors on the 5nm process node at Taiwan Semiconductor Manufacturing Company (TSMC) but is moving to the newer, mature 4nm process node that NVIDIA is using to build its Grace CPU and "Hopper" H100 and H200 AI GPUs.
Inside, the new Graviton4 has 96 cores based on Arm's "Demeter" Neoverse V2 core (which is based on the Armv9 architecture), which is a 50% boost in cores over Graviton3 which had 64 cores.
Intel Xeon 'Emerald Rapids' Xeon 8551C, 8558P CPUs teased... AMD EPYC will smash them easy
We first started hearing about Intel's next-gen Xeon "Emerald Rapids" CPUs back in August 2021; fast-forward two years, and we're still not seeing them in the wild, with new rumors teasing Intel's upcoming Xeon Platinum 8551C and Xeon Platinum 8558P "Emeralds Rapids" processors.
Intel's new Xeon Platinum 8551C and Xeon Platinum 8558P "Emeralds Rapids" processors both feature 48 cores and 96 threads of CPU power, recently tested on the CentOS Linux OS and 256GB of DDR5 memory. Geekbench results see both the new Emerald Rapids CPUs at up to 3.17GHz during their testing, with the 8551C featuring a 2.9GHz base clock and the 8558P featuring a 2.7GHz base clock.
But... AMD has released its consumer-focused HEDT processors with the new Ryzen Threadripper 7000 series featuring up to 64 cores and 128 threads on the desktop, not an Xeon-based CPU for servers. There's also the Ryzen Threadripper 7970X processor with 32 cores and 64 threads that comes close to the multi-core performance in Geekbench against the Xeon Platinum 8551C.