Processors - Page 68

Intel's next-gen Sapphire Rapid-X architecture isn't far away with the new HEDT platform, which will be made on the new Intel 7 node.

According to the very latest rumors from Tom @ Moore's Law is Dead, we're to expect up to 56 cores and 112 threads of CPU power and DDR5 RAM + PCIe 5.0 support. We are to expect an edition of Sapphire Rapids that will rock super-fast HBM2e memory and a nifty new Ponte Vecchio GPU for the new Aurora supercomputer.

The new rumor from MLID teases that Intel is actually going to drop the Core-X branding with Sapphire Rapids, and launch it as the new Xeon Workstation family of processors. With up to 56 cores, 112 threads, 8-channel DDR5 support, 350W of TDP, and an insane 112 x PCIe 5.0 lanes... Intel is back.

MediaTek has just unveiled its new 5G-powered Dimensity 9000 SoC which features 8 cores of CPU power, an ARM Mali-G710 MC10 GPU, ray tracing abilities, 5G connectivity, and it's made on TSMC's new 4nm process node.

The new MediaTek Dimensity 9000 SoC has 1 x Cortex-X2 at 3.05GHz, 3 x Cortex-A710 at 2.85GHz, and 4 x Cortex-A510 at 1.8GHz on the CPU side of things -- joined by 8MB of L3 cache and 6MB of system-wide cache. This some high-end CPU grunt here, which should see MediaTek better battling Qualcomm, Samsung and Apple in the mobile high-end SoC battle.

We have LPDDR5X memory clocked at up to 7500 MT/s although JEDEC has support for up to 8533 MT/s, but even at the higher LPDDR5X-7500 memory is 17% faster than the LPDDR5-6400 standard.

Intel may have just launched the new 12th Gen Core "Alder Lake" CPUs but the new 13th Gen Core "Raptor Lake" CPUs will launch next year -- and then in 2023 we'll see the 14th Gen Core "Meteor Lake" CPUs which we have a closer look at right now.

CNET recently walked through Intel's new chipmaking facility in Arizona, where they snapped some beautiful photos of Intel's new 14th Gen Core "Meteor Lake" mobile processors. Inside of the Intel Fab 42 facility in Chandler, Arizona, CNET snapped some photos of the new Intel Meteor Lake-M processor, which should debut in a couple of years time.

Intel's next-gen Meteor Lake-M should arrive with a 5W to 15W TDP, with multiple tiles on the package with Intel Foveros technology and could have up to 3 tiles. This should be split between the Computer Die (which would be the biggest tile on the photo), the SoC-LP for I/O operations, and then a graphics die. The GPU side of things would see something between 96 and 192 Execution Units (EUs or GPU cores).

Intel has detailed its next-gen Xeon platform "Sapphire Rapids" which will deliver up to 56 cores and 112 threads per CPU, with multiple Xeon "Sapphire Rapids" CPUs per socket -- on top of that DDR5 RAM and PCIe 5.0 technology support.

Intel's new Xeon "Sapphire Rapids" CPU has 4 stacks of 16GB HBM2e memory, for a total of 64GB HBM2e through Xeon Scalable. The new Sapphire Rapids CPUs are built on Intel's new Eagle Stream platform, unlocking up to 56 cores and 112 threads with up to 350W TDPs.

On the GPU side of things Intel is using its new Ponte Vecchio HPC GPU, offering 4MB of L1 cache and 144MB of L2 cache per GPU tile. In total, there is 408MB of L2 cache while the GPU is going to use multiple fabrication notes: Intel 7 for the base tile, TSMC N7 for Xe-Link, and TSMC N5 for the computing tile.

Can you believe it has been 50 years since Intel released the Intel 4004 processor, the world's first commercially available microprocessor. 50 years from that day, we have Intel's new 12th Gen Core "Alder Lake" CPUs that have just launched.

The Intel 4004 processor in November 1971 and acted as the "brains" the virtually the foundation of technology as we know it: from your smartphone to your desktop PC, right up into the cloud. Microprocessors opened up the world to what it is today, and without them we'd still be using paper cups and string... well, you get what I mean.

Maybe Pat Gelsinger, Intel CEO will explain it better: "This year marks the 50th anniversary of the 4004 chip. Think of how much we've accomplished in the past half-century. This is a sacred moment for technology. This is what made computing really take off!"

We're hearing more about AMD's new mobile powerhouse "Raphael-H" processors, after previous rumors teased some details we're now hearing some more "concrete" leaks.

According to leaker Greymon55, AMD's new Rembrandt-H might be limited to just 8 cores but things will be ramping up considerably with Phoenix-H which will be made using the new Zen 4 architecture and on the new 5nm process node. AMD's codename Phoenix-H will have up to 8 cores, 16 threads, and a TDP of less than 40W.

But after that, AMD's new codename Raphael-H processor will use the same Zen 4 architecture and 5nm process node, but ramp up core counts considerably: up to 16 cores and 32 threads with a TDP of over 45W. AMD's new "Raphael" platform will be on both the desktop and laptop, upgrading to the Zen 4 architecture, new AM5 platform, DDR5 support, an integrated Navi 2-based GPU and so much more.

AMD announced its next-gen Zen 4-based EPYC "Genoa" CPUs today at its Accelerated Data Center event, as well as the world's first MCM-based "Aldebaran" GPU that will power AMD's new Instinct MI250X and MI250 accelerators.

AMD also had a huge announcement in that they have secured Meta, the company formerly known as Facebook, as a new customer. Meta will be using AMD EPYC CPUs inside of its data centers, and from that news alone AMD shares skyrocketed over 11%.

The company showed off its new 3rd Gen EPYC CPUs that have AMD's exciting new V-Cache technology, with the EPYC "Genoa" CPUs launching with up to 96 cores in Q1 2022. AMD EPYC "Genoa" will be followed up with AMD "Bergamo" that will have the newer Zen 4c architecture and a hugely boosted CPU core count of 128 cores and 256 threads.

AMD has just confirmed its new EPYC "Genoa" and EPYC "Bergamo" CPUs at its Accelerated Data Center event -- where we were just introduced to the NVIDIA-destroying Instinct MI250X accelerator with its MCM GPU and 128GB of HBM2e memory.

The new AMD EPYC codename "Genoa" CPU will be based on the new Zen4 processor architecture from AMD, and will have a huge 96 cores -- just as the rumors said it would -- and support PCI 5.0 technology, the new DDR5 memory standard, and will be AMD's first platform with CXL interface support.

AMD's new EPYC "Genoa" is going to be based on the Zen 4 architecture, while the higher-end EPYC "Bergamo" that we've also been hearing lots about, will arrive on the "Zen 4c" core. EPYC Genoa will have 96 cores, while EPYC Bergamo ramps things up to a huge 128 processor cores -- the first of its kind, from AMD, again.

Intel might have just launched its new 12th Gen Core "Alder Lake" CPUs but even in the months leading up we were enjoying leaks and rumors of the 13th Gen Core "Raptor Lake" CPUs, and today we have a new tease.

The new Intel 13th Gen Core "Raptor Lake" CPUs will arrive in 2022, and according to Tom @ Moore's Law is Dead, Raptor Lake will also support DDR5 and DDR4 memory, just like the just-launched 12th Gen Core "Alder Lake" CPUs and the new Z690 chipset.

We will see Intel launch its new 700 series chipset with the flagship Z790 chipset lighting up next-gen Raptor Lake-ready motherboards in 2022. If you've got some good DDR4 memory now and didn't want to upgrade to Alder Lake, you should be fine when Raptor Lake launches and not need to buy new DDR5 RAM.

Apple's next-generation iPhone will use a new chip based on the 4nm according to new rumors, down from the 5nm chip inside of the iPhone 13 and iPhone 12 families of smartphones.

The new iPhone 13 uses an A15 Bionic chip on the new 5nm process from TSMC, while the iPhone 12 used an A14 Bionic chip was also made on the 5nm node at TSMC. But the next-gen iPhone 14, which is meant to be a "complete design" will also have a nifty new 4nm chip in the new A16 Bionic.

The Information reports that Apple and TSMC are having trouble making a 3nm chip, which might result in the new iPhone 14 rolling out with a "4nm" process instead. We already know that Apple has scooped up all of TSMC's production capacity fo its next-gen 3nm process, which should find a new home inside of the iPhone 15 and future releases of M1 chips for next-gen MacBook laptops and Mac desktops.

Qualcomm hosts its annual Snapdragon Tech Summit in early December, where I now miss attending the in-person events held in Maui, Hawaii. But this year, we're to expect the next-gen Snapdragon 898 SoC to be unveiled.

The next-gen Qualcomm Snapdragon 898 should be unveiled on day one of the Snapdragon Tech Summit, on November 30, with the event running through to December 2 -- as it always did -- as a 3-day event. The new Snapdragon 898 will reportedly be mass-produced on Samsung's new 4nm process, and not TSMC's new 4nm process.

As for the performance upgrades, we should expect the new Snapdragon 898 to offer around 20% more performance over the current-gen flagship Snapdragon 888. Apple is on the scene in a big way with its new in-house M1 Max and M1 Pro SoCs that are offering tremendous performance in a mobile chip, made on TSMC's fresh new 5nm node.

Apple has made a very big splash into the industry by diving head-first into making its new in-house M1 Max and M1 Pro SoCs a force to be reckoned with, even against the mighty Intel -- but what comes next is going to be even bigger.

The latest rumor has Apple's next-gen SoC packing 40 CPU cores on the next-gen 3nm process, furthermore, Apple is working with TSMC to make its next-gen on an enhanced version of the 5nm process. The new chips will reportedly feature 2 dies that will let Apple cram in many more cores, making next-gen MacBook Pro laptops and Mac desktops much, much more powerful.

The new 3nm chips can have up to 4 dies that would allow for up to 40 computing cores, compared to the current-gen M1 Max with a 10-core CPU. The current-gen Apple M1 Max features a 10-core CPU, up to a 32-core GPU, up to 64GB of RAM, packs a huge 57 billion transistors, and is made on TSMC's new 5nm process.

AMD has had success after success, milestone after milestone, and record after record broken with its Zen CPU family of products: Ryzen, Threadripper, and EPYC... now, they've reached a new CPOU market share milestone -- something AMD hasn't been near since the end of 2006.

According to research analyst firm Market Research, AMD now has 24.6% of the CPU market share which is a position the company hasn't been in since Q4 2006... 15 years ago. This is the second-highest CPU market share that AMD has held where in Q4 2006 the company had 25.3% against Intel.

AMD also reached a new all-time high in notebook x86 unit share with 22%, a gain of 1.8% year-over-year. AMD has gone from stride to stride with clawing back large amounts of CPU market share from Intel with its Zen-based Ryzen CPUs for the desktop being a major success with everyone from entry-level users to gamers and enthusiasts.

AMD didn't make too much of a fuss when its 4700S processor launched, but now Minisforum is using the AMD 4700S processor inside of its new CR50 system.

The new Minisforum CR50 packs the AMD 4700S 8-core processor kit, AMD Radeon RX 550 graphics, and 16GB of GDDR6 memory with options for either 16GB RAM + 256GB SSD or 16GB of RAM + 512GB SSD. The 8-core AMD Zen 2-based CPU has a max boost clock of up to 4.0GHz, with 512KB of L1 cache, 4MB of L2 cache, and 8MB of L3 cache.

Minisforum's new CR50 mini PC is pretty damn small, measuring in at 205mm long, 192mm wide, and 86mm tall. There's plenty of connectivity but only HDMI and DVI display outputs, no DisplayPort here... but you do have USB 3.2 and USB 2.0 ports, as well as the usual GbE ethernet connection and 3.5mm audio jacks.

Intel's new 12th Gen Core "Alder Lake" CPUs are about to be unleashed onto the world, with the flagship Core i9-12900K processor already pushed to record-breaking heights: the land of 8GHz.

Intel's new Core i9-12900K is the flagship Alder Lake-S processor that was overclocked to 8GHz by extreme overclocked HiCookie, who used the new AORUS Z690 Tachyon motherboard. DDR5 memory was used, of course, also overclocked to a blistering DDR5-8300 on looser 52-52-52-100-127-2 timings.

You're not going to be overclocking Intel's new Core i9-12900K processor to 8000MHz on the stock cooler, not even a huge 360mm AIO liquid cooler or a custom water loop... HiCookie used exotic LN2 cooling, used by extreme overclockers. Still, the Core i9-12900K pushed to 8GHz says a lot about Intel's shift to its new Intel 7 node with Alder Lake as the previous-gen Core i9-11900K was never pushed to 8GHz.

AMD's next-gen Zen 4 Dense (Zen 4D) has been teased from leaker Tom at Moore's Law is Dead, with Zen 4D packing a "completely redesigned cache system, stripped down features, and reduced target clockspeeds for lower power consumption and increased density".

Alrighty... all of that, but it also "trades away some single core performance for massively increased MT in a similar die space to Zen 4". OK, you have my attention now... especially with the added tease that Zen 4D-based processors would have 16 cores per chiplet and DDR5 RAM support.

Tom says that AMD could make a 2 x 16C Zen 4D-based "R9 7965WX" processor for the AM5 processor in Q1 2023 as an "alternative to the flagship 16C Raphael" processor. But the big question is... if you didn't already know the answer: what is this new CPU going to be targeted towards? Intel's new Hybrid CPU technology.

AMD is sitting on its hands right now riding out the wave of Intel's about-to-launch 12th Gen Core "Alder Lake" CPUs and the new Z690 chipset, with its own next-gen CPUs dropping in 2022.

According to Yuri "1usmus" Bubliy, the creator of the DRAM Calculator and Clock Tuner for Ryzen says that AMD might be releasing new Ryzen CPUs in late January 2022. Bubliy says he's working on Project Hydra, a new tool that will support the current Ryzen 5000 series, and next-gen Ryzen 6000 series CPUs.

The second he said that there were instantly tweets asking what the hell is going on with Hydra and the new Ryzen 6000 series CPUs that no one else knows about. What we do know about the new CPUs is that they should be Zen 3-powered CPUs with new 3D V-Cache technology.

MSI has revealed quite a lot of interesting factoids on Intel's new Alder Lake platform, including DDR5 and PCIe 5.0 teases -- but now the company has revealed there are not 1 but 2 different dies of Intel's new 12th Gen Core "Alder Lake-S" desktop CPUs.

During its recent Innovation Event, Intel confirmed there are at least 60 different SKUs of their Alder Lake silicon in the works, and then during the MSI Insider stream the company confirmed that desktop Intel Alder Lake CPUs actually ship in 2 different variants.

The first is with 8 Golden Cove and 8 Gracemont CPU cores -- while another ships with just 6 Golden Cove CPU cores. Intel uses the Golden Cove CPU cores as its Performance cores (or P-core) while the Gracemont CPU cores are the Efficient CPU cores (E-cores). Intel hasn't used the 6P+0E die just yet and should arrive in early 2022 when we see Intel fill out its Alder Lake family of CPUs (non-K Cores series CPUs, as well as Pentium CPUs).

Intel's new 12th Gen Core "Alder Lake" CPUs aren't available to the public for another few days yet, but Newegg has shipped batches of the Alder Lake CPUs to some customers, some of which have been contacted to ask for them to not break the embargo from Intel.

Newegg might have shipped some of Intel's new 12th Gen Core "Alder Lake" CPUs but they will be useless, because they'll need a new Z690-based motherboard which aren't available yet. So while Newegg has shipped some Alder Lake CPUs, they didn't send Z690 motherboards or DDR5 RAM.

Some of those new Intel Alder Lake CPU owners posted on Reddit, with "Hyphnx" saying: "Getting my 12900K tomorrow, what can't I post? Just got a phone call from Newegg and the lady said I can't post anything until 11/4?" while "NoveltyNetwork" said: "Just got the same call, and mine just came in about an hour ago. The packaging on the 12900k looks awesome! The 12700k looks a bit mediocre". "discovet11" said: "Just got a voicemail from a lady as well. Told me not to post anything until 11/4 "to avoid any issues".

Intel might have just launched its 12th Gen Core "Alder Lake" CPUs but AMD is stealing headlines with its from-the-future EPYC "Turin" CPUs which will be based on the Zen 5 architecture.

AMD's next-next-gen EPYC "Turin" CPUs on the Zen 5 architecture will boast up to 256 cores and 512 threads, with all of that silicon having up to 600W of power at its hands. AMD will be using a stacked 3D chiplet design with EPYC "Turin", an evolution of what we'll see in the new EPYC "Milan-X" CPUs later this year.

AMD's upcoming EPYC "Genoa" CPUs will have up to 96 cores and 192 threads on the Zen 4 architecture and supports huge 12-channel DDR5 memory. The leap from AMD's EPYC "Genoa" CPUs to the next-gen EPYC "Turin" will be equally as large: Turin will use the Zen 5 architecture, could boast PCIe 6.0 connectivity, and an insane 256 cores and 256 threads on a single CPU.