Processors - Page 29

Intel is weeks away from launching its new Core Ultra 200 series "Arrow Lake-S" processors, with recent rumors suggesting the Arrow Lake Refresh has been nixed and that Nova Lake will succeed Arrow Lake... and now on top of that, rumor has it that Razer Lake-S will succeed Nova Lake-S in 2026+.

It looks like Intel's upcoming Core Ultra 200 series "Arrow Lake-S" desktop processors will be on the market much longer than expected, with the Arrow Lake-S Refresh canceled, it's all-systems-go into Nova Lake, and then into the arms of Razer Lake-S.

LGA 1700 has been used for Alder Lake-S, Raptor Lake-S, and the Raptor Lake-S Refresh CPUs but Intel will debut its new LGA 1851 socket with Arrow Lake-S (with LGA 1851 meant to stick around for Arrow Lake-S Refresh). But now... well, now there's the issue of whether Intel continues on with LGA 1851 for Nova Lake-S and Razer Lake-S, but personally, I think changing sockets right after LGA 1851 is going to set off alarm bells for consumers.

Intel has reportedly nixed its upcoming Arrow Lake-S Refresh, where rumor has it that Nova Lake will now succeed the upcoming Core Ultra 200 series "Arrow Lake-S" desktop CPUs.

In a new post on Chiphell and X, leaker "Panzerlied" said that the Arrow Lake Refresh has been "canceled" adding that Arrow Lake-S will "battle for two years". Intel appears confident that its upcoming Core Ultra 200 series "Arrow Lake-S" processors that it will push out its next-gen CPU launch until Nova Lake.

Arrow Lake-S Refresh would've been compatible with the new LGA 1851 socket, but we can expect Nova Lake to slot right into the new LGA 1851 socket, but things could change (I doubt it, changing sockets from LGA 1700 to LGA 1851 and then again in two years would look HORRIBLE against AM5 and its many years of life so far, and into the future).

Apple has debuted its upgraded A18 Pro SoC inside of its new iPhone 16 family of handsets -- inside of the higher-end iPhone 16 Pro and iPhone 16 Pro Max handsets -- and now we've got some benchmarks that show it's the fastest mobile SoC in the world.

The new Apple A18 Pro is fabbed on TSMC's newer 3nm process node (N3B) with reviewer "Geekerwan" posting some ranking charts showing how the new A18 Pro fares against some of the most powerful mobile SoCs on the market including the Qualcomm Snapdragon 8 Gen3, Dimensity 9300, and Apple's more powerful M2, and M4 SoCs inside of new MacBook laptops.

In CPU rankings, the new A18 beats the previous-gen A17 Pro, while the new A18 Pro beats out the Snapdragon 8 Gen3 and Dimensity 9300, while losing to the M2 inside of the iPad Pro, and the M4 (3+6) and M4 (4+6) by a larger margin. Still, it's impressive to see the new A18 Pro getting closer to the much higher-end M2 and M4 processors.

Intel's new Core Ultra 7 268V "Lunar Lake" CPU has been tested ahead of laptops hitting the market later this month, with the 8-core, 8-thread CPU tested against AMD's new Strix Point APU and Apple's new M3 processor.

Inside, the new Intel Core Ultra 200V series "Lunar Lake" CPUs feature new Lion Cove P-Cores and Skymont E-Cores, with either 16GB or 32GB of on-package memory that sits next to the chip itself. In some early testing, we've got some numbers from a Vietnamese YouTuber that has tested a new Dell XPS 13 laptop with the Intel Core Ultra 7 268V processor, 32GB of LPDDR5X-8533 memory, and Windows 11 Home.

Cinebench R23 single-threaded performance shows that the new Intel Core Ultra 7 268V "Lunar Lake" processor beating out the AMD Ryzen AI 9 HX 370 "Strix Point" APU, and the Apple M3. Interestingly, the Core Ultra 7 268V processor using 15W of power, the Ryzen AI 9 HX 370 APU using 21W of power, while Apple blows them both away with the M3 using just 6W of power.

Intel's new Core Ultra 200 series "Arrow Lake" desktop CPUs will be here soon, with their retail packaging now leaked showing us the refreshed look that Intel has for Arrow Lake (and the new Core Ultra naming scheme for the desktop). Check them out:

It looks like Intel is going with a darker blue rectangular box with dots as accents, with the "ULTRA" branding in a smaller font than the "Intel CORE" above it. This is similar to previous-gen Core boxes in retail form, except we've moved away from the Core i3, Core i5, Core i7, and Core i9 into the Core Ultra 5, Core Ultra 7, and Core Ultra 9 processors for Arrow Lake.

Intel will debut its Arrow Lake-S desktop CPU family with the flagship Core Ultra 9 285K processor, with 24 cores and 24 threads (no Hyper-Threading) and CPU boost clocks of up to 5.5GHz. The Core Ultra 9 285K will feature 8 P-Cores based on the Lion Cove architecture, and 16 E-Cores based on the Skymont architecture.

Intel is preparing new Core Ultra 200HX series "Arrow Lake-HX" enthusiast laptop CPUs, with new leaks pointing to the flagship Core Ultra 9 285HX processor with up to 24 cores and up to 5.5GHz boost clocks.

In a new post on X, leaker "Jaykihn" posted details on Intel's fleet of Arrow Lake-HX mobile processors for new enthusiast gaming laptops that I'm sure we'll see debut at CES 2025 in January 2025. We are to expect 6 different SKUs of the Core Ultra 200HX series CPUs, which will share a similar die to the desktop Arrow Lake-S processors.

We can expect the following Arrow Lake-HX processors:

AMD only just released its Ryzen AI 300 series "Strix Point" APUs, but with a heavily beefed-up "Strix Halo" teased in the rumor mill for months now... but now we're hearing about a new Ryzen AI Max naming scheme for Strix Halo.

The company is reportedly working on new Strix Halo APUs that will be split into three newly-rumored SKUs: the Ryzen AI Max+ 395, Ryzen AI Max 390, and Ryzen AI Max 385 each with differing CPU and GPU core counts.

Each of the new Strix Halo APUs will supported up to a monster 96GB of DDR5 RAM, which completely destroys the hardware limit of just 32GB of on-package memory with Intel's soon-to-be-here Core Ultra 200V series "Lunar Lake" processors.

Intel's new Core Ultra 200V series "Lunar Lake" CPUs are nearly here, launching on September 3, they're slowly making their way into laptops across the world and we've got some early results that see it beating Apple's new M3 and M2 MacBook laptops.

Lenovo has teased its upcoming Yoga Slim 7i Aura Edition laptop, which features the Intel Core Ultra 7 258V "Lunar Lake" CPU and has been benched against the M2 and M3 MacBook laptops from Apple. Lenovo is proud to have all-day, virtually 24-hour battery life thanks to the impressive new feats of the Lunar Lake CPU from Intel.

The company tested Local Video Playback abilities of the Lunar Lake CPU-powered laptop, with 150 nits of brightness while looping a video sample based on the H.264 (1920 x 1080 at 24FPS) format. Intel's Core Ultra 7 258V CPU features 32GB of on-package RAM, with the system running Windows 11 with the 24H2 update. Lenovo adds that this is an engineering sample, and that results could change a little (probably for the better).

AMD is reportedly cooking up some new Ryzen 200 series "Hawk Point Refresh" APUs to fight a future release of Core 200 series "Raptor Lake Refresh" CPUs from Intel later this year.

In a new post by leaker Golden Pig Upgrade, we're learning that AMD is dropping the "AI" naming from this APU, as the company just launched its new Zen 5-powered Ryzen AI 300 series "Strix Point" APU with an on-board NPU.

Golden Pig Upgrade said: "In order to cope with the release of Raptor Lake Refresh, also known as the Core 200 series (not Core Ultra) at the end of the year, AMD plans to rename the current Hawk Point, also known as the 8040 series, to the Ryzen 200 series (not Ryzen AI) next year. I think our A is still conservative and should be renamed Ryzen 300".

Intel's Arrow Lake processors are presumably around the corner, with the latest rumors pointing to a release of the new desktop CPUs sometime in late October. 

Ahead of the highly anticipated launch reports have come out that Intel's next-generation of desktop CPUs will come with a significant bump in DDR5 memory speeds, with WCCFTech reporting Core Ultra 200 series CPUs will come with support for 8000-10,000 MT/s with CUDIMM. For those that don't know, CUDIMM or Clocked Unbuffered Dual In-Line Memory Module, is the latest form of DDR5 memory that only came out this year. It was actually showcased at Computex 2024. 

The leak about a bump in DDR5 memory speeds comes from MebiuW from Weibo, who wrote that if this performance increase is true, Intel would have widened the memory capability gap between itself and AMD, which is currently struggling to reliably hit 7,000 MT/s. Intel's previous generation, called Raptor Lake, had a difficult job of reaching 8,000 MT/s, which means if these rumors are true, Intel would have made a significant bump in memory speeds with its Intel Core Ultra 200 series. 

Apple is now having its A16 processor made at TSMC's new fab in Arizona, under the same 4nm N4P process node that TSMC's Taiwan factories are using for the best quality and performance.

We have been hearing good things about TSMC's new Fab 21 facility in Arizona, and now that it's pumping out Apple A16 processors, it's another good sign... and fantastic for the United States. Apple is an American company, with Taiwanese giant TSMC opening up a fab on US soil, and now fabbing chips for the iPhone 14.

TSMC is expected to ramp up production at its Fab 21 plant in Arizona over the coming months, with the second stage of its first phase completed, full-scale manufacturing can begin in the first half of 2025. It seems Apple is confident enough with TSMC's ability to make A16 chips at its Arizona fab, because if there were any yield or performance issues, Apple of all companies wouldn't be using them.

AMD's new Zen 5-based Ryzen 9 9700X has been overclocked to 6.3GHz using exotic LN2 cooling, beating out the Intel Core i9-14900KF processor overclocked to a whopping 7.1GHz.

The huge achievement was reached by overclocker Skatterbencher, using the new Ryzen 9 9700X to break the OCCT AVX benchmark with a world record 269.35 points, with the 9700X overclocked using LN2 cooling and hitting 6318MHz (6.31GHz).

Skatterbencher tweaked the hell out of the Ryzen 9 9700X processor, where instead of using manual multiplier and voltages, he used a combination of BLCK overclocking, Precision Boost Overdrive (PBO), AMD's Curve Optimizer (with a positive offset), and Curve Shaper to enable Ryzen's Precision Boost algorithm to run at over 6GHz using liquid nitrogen cooling.

Japan is stepping ahead of the world in the use of synthetic diamonds to power next-gen semiconductors, offering performance in orders of magnitude beyond current materials, and it's getting closer to commercial viability thanks to advances coming out of Japan.

These new semiconductors are made from synthetic diamond, which has been referred to as the "ultimate semiconductor material" thanks to its beyond-impressive thermal conductivity and electrical strength. Diamond is fantastic to power semiconductors, as it is around 33x stronger in electric strength as an insulator as silicon.

Diamond power semiconductors can also operate in environments around 5x as hot, and theoretically have the capacity to handle around 50,000x as much electricity. Yes... 50,000x the electricity through diamond power semiconductors, all advancing from Japan. Silicon Carbide (SiC) and gallium nitride (GaN) have received attention recently as the next-gen semiconductor substrates, but the performance on the table with diamond is far better.

Intel is reportedly working on a next-generation Cobra Cove x86 architecture, which will power future-gen processors with higher performance and power efficiency.

We've been hearing about Royal Core, a project started by chip legend Jim Keller, which was meant to power next-gen Nova Lake and Beast Lake, with Beast Lake Next thrown into that mix but reportedly canceled. Intel is starting the architectural change journey with its upcoming Core Ultra 200 series "Arrow Lake" CPUs, with a focus on single-core performance with the removal of Hyper-Threading.

Royal Cove has been rumored with an interesting introduction: Rentable Units, in its first generation lacking Hyper-Threading, but Royal Cove 1.1 was meant to feature a new method of Hyper-Threading which splits the P-Cores into two smaller cores, while Beast Lake Next used Royal Core 2.0 that was rumored to feature up to 4 threads per P-Core, but these were just rumors... and now we're hearing about Cobra Cove. Let's go.

"In our last quarter, I think data center was over 50% of our revenue," AMD CEO Lisa Su said at the recent Goldman Sachs Communacopia And Technology Conference. Adding, "So we really are a data center first company." Although this might sound like a shift in focus for AMD, with the generative AI boom in full swing, it's par for the course for chipmakers feeding the insatiable AI beast.

Looking at AMD's most recent earnings, data center revenue, which includes server-specific EPYC CPUs and other GPU hardware, brought in roughly double the revenue of AMD's client and gaming business. Lisa Su added that AMD's data center-first approach extends to cloud, edge, and client - as seen in the latest Ryzen AI chips for mobile devices.

It's a similar story at NVIDIA, which has seen unprecedented growth and expansion in recent years due to its data center and AI software business.

Intel is providing a full refund or swapping out affected 13th Gen Core and 14th Gen Core CPUs with the Core i9-14900K, with Intel's hero act possibly starting.

Intel had flaws in its RMA process, seeing the company reject "valid" RMA requests just weeks ago, but now, there has been multiple people posting on social media that have worked the system, leveraging the flaw in Intel's RMA process and benefitting big time.

X user "Pk1" posted that he sent a fully functional "never game me an issue" almost 2-year-old Core i9-13900K, to which he received a $599 check. He didn't need to provide any proof of instability, and Intel even paid for the shipping. In full, Pk1 said: "Well, it's 100% official. I received a check for $599 from Intel today for my fully functional, never gave me an issue, almost 2yrs old 13900k. They required zero proof of instability and even paid for the shipping. I wish this happened every generation!!! Lol"

AMD has just announced its new 10.10 "Advancing AI 2024" event, where we will be introduced to the EPYC "Turin" CPU, Instinct MI325X AI accelerator, and Ryzen AI 300 PRO series CPUs.

This could end up being one of AMD's biggest launches of the year, after its huge Zen 5 reveal over the last couple of months, we're looking at a new server CPU, new AI accelerator, and new PRO series Strix Point APUs. That's a lot of tech, all while providing the chips it does to millions of systems and devices worldwide, and Sony's just-announced PS5 Pro console. AMD isn't finished, and then we have CES 2025 where we'll be introduced to the new "Strix Halo" and "Krackan Point" APUs.

AMD explained on its website: "Today, AMD announced"Advancing AI 2024,"an in-person and livestreamed event on October 10, 2024 to showcase the next-generation AMD Instinct™ accelerators and 5th Gen AMD EPYC™ server processors, as well as Networking and AI PC updates, in addition to highlighting the Company's growing AI solutions ecosystem. AMD executives and AI ecosystems partners, customers and developers will join Chair and CEO Dr. Lisa Su to discuss how AMD products and software are reshaping the AI and high-performance computing landscape".

AMD has confirmed it will have its new "Krackan" APU launching in early 2025, aiming at the $799 notebook market ready for Copilot+ in 2025.

The big difference between the just-launched Ryzen AI 300 series "Strix Point" APUs and the new "Krackan" (also rumored as "Kraken Point") is that it will drop down to 8 cores of Zen 5 (4 x Zen 5 + 4 x Zen 5c) and a heavily cut-down 8 CUs of RDNA 3.5 GPU (compared to 16 CUs of RDNA 3.5 on Strix Point, and 40 CUs of RDNA 3.5 on Strix Halo).

AMD's upcoming Krackan APU will feature the same Zen 5 CPU + RDNA 3.5 + XDNA 2 NPU, so we have Copilot+ ready for 2025, with enough TOPS (45-50 TOPS) for AI workloads of the future (and to compete against Intel Lunar Lake and Qualcomm Snapdragon X).

AMD is reportedly cooking up a next-generation Ryzen Z2 Extreme SoC, something revealed during a recent join Q&A session at IFA 2024 between AMD and Microsoft.

At the event, Digital Trends was in attendance when AMD explained that AMD is aiming for an "early 2025" release for its new Z2 Extreme chip. This is incredibly exciting news as AMD has powered most of the best gaming handhelds on the market with its Ryzen Z1 and Z1 Extreme processors, inside of the likes of the ASUS ROG Ally.

We should expect some major upgrades from the successors to most gaming handhelds in 2025, powered by AMD's upcoming Ryzen Z1 Extreme. AMD's Jack Huynh talked about wanting to play Black Myth: Wukong for 3 hours on a handheld, not just the 45 minutes we get from current-gen handhelds.

Intel has launched its new Core Ultra 200V series "Lunar Lake" CPUs with new laptops powered by Lunar Lake not hitting the market until September 27... and now we've got some simulated benchmark numbers from Intel's next-gen consumer desktop CPU: Arrow Lake.

The new Core Ultra 200 series "Arrow Lake" (the 200 series = desktop, while 200V series = laptop) has been put through a "simulated" benchmark by leaker Moore's Law is Dead. These numbers from the Core Ultra 9 285K are at its best, with the ideal IPC improvement over Raptor Lake, and it beats the Core i9-14900K but loses to the Ryzen 7 7800X3D with the new EXPO update. MLID assumes 5.4GHz for the 9800X3D (assume 2% IPC gain in gaming) while perfect 5.7GHz and "perfect gaming IPC uplift" for the Core Ultra 9 285K chip.

In some simulated benchmarks of the upcoming Intel Core Ultra 9 285K processor against AMD's upcoming Ryzen 7 9800X3D, the new Arrow Lake-S desktop CPU loses to the Zen 4-based Ryzen 7 7800X3D, and loses with an even bigger margin against the Zen 5-based Ryzen 7 9800X3D.