CPU, APU & Chipsets News - Page 4
Intel has now better detailed its upcoming ARC series GPUs, with its gaming-focused ARC Alchemist DG2 GPU confirmed to rock the next-gen DisplayPort 2.0 standard.
Intel's first-gen ARC graphics cards are codenamed "Alchemist" and will support DisplayPort 2.0, which supports some insane resolutions and refresh rates: we're talking up to 4K 240Hz and 8K 120Hz. The current HDMI 2.1 standard which was only supported by AMD's new RDNA 2 and NVIDIA's new Ampere GPU architectures, supports only 4K 144Hz and 8K 60Hz maximum.
The news of DisplayPort 2.0 support comes from something Phoronix spotted, with the DG2 GPU family of cards mentioned in patch notes for the i915 kernel display driver. This isn't instant DP2.0 support, but rather that engineers are working on it right now -- and that when Intel's new ARC Alchemist launches, the cards will have support for the new DisplayPort 2.0 standard.
Oh Tom, man I love you -- spoiling all of the future fun for Intel ahead of time, after being proven right during Intel Architecture Day 2021 multiple times. Now, for some juicy details on Intel's "Royal Core" project -- the AMD Zen 5 architecture killer.
I've been hearing from some of my sources for the last 18+ months or so that what Intel has coming will "destroy AMD" in the future, but now Tom from Moore's Law is Dead is providing much more details on that AMD killer -- the Royal Core Project.
Intel hired Jim Keller, the mastermind behind AMD's return-to-form Zen CPU architecture, leaving AMD for Tesla and then Tesla for Intel, and then Intel for something new. In that time at Intel, it appears Keller designed the Royal Core Project -- this is the architecture that moves past Core, the largest upgrade in CPU architectures for Intel... ever.
Intel has just wrapped up its exciting new Architecture Day 2021 event, with details on their future CPUs and GPUs in the oven right now -- but Tom over at Moore's Law is Dead has decided to drop a new video leaking out some new future Intel goodies.
We're finding out details on Intel's new Emerald Rapids CPU architecture, which will succeed the upcoming Sapphire Rapids CPU architecture. Intel's next-gen Emerald Rapids would launch with up to 64 cores at up to a whopping 350W and it'll be made on Intel's upcoming Intel 7 node.
But more surprisingly, Emerald Lake would roll out with an insane 80 PCIe 5.0 lanes and support for DDR5-5600 memory. 80 PCIe 5.0 lanes just sounds out-of-this-world to me, or anyone right now, as you're running out of PCIe 3.0 or PCIe 4.0 lanes on a consumer motherboard once you have a x16 graphics card and anything else that requires PCIe lanes (think NVMe M.2 SSDs, etc).
Intel's next-gen Raptor Lake-S processors will be coming in late-2022, and now we have some more juicy rumors to feed you in the form of the flagship Core i9-13900K processor.
The flagship Core i9-13900K processor will have 32 threads with a configuration of 8 cores + 16 cores, for a total of 24 threads of CPU power. Intel will be tapping the same hybrid architecture that will ship inside of the next-gen Alder Lake-S processors that will drop in a couple of months' time, but Raptor Lake will have an updated high-performance core codenamed Raptor Cove.
Alder Lake will ship with new high-performance "Golden Cove" cores and high-efficiency "Gracemont" cores, while the next-next-gen Raptor Lake CPUs will have new high-performance core codenamed Raptor Cove. Not to be confused with Raptor Lake, which is the codename of the family of CPUs -- but the tweaked high-performance core "Raptor Cove" will be the big difference between this year's Alder Lake and next year's Raptor Lake chips.
You might have heard about the GIGABYTE hack and leak of secret documents... well, it looks like we're getting some information from that hack in the form of AMD's next-gen EPYC "Genoa" CPUs.
The next-gen EPYC "Genoa" CPUs will be based on AMD's next-gen Zen 4 microarchitecture, with new details on the AMD SP5 and AM5 platforms -- the first that support the Zen 4 architecture for servers, and consumers respectively.
The documents in question are dated July 2021, so they're still new -- and talk about motherboard thermal and power compatibility guidelines. Not only that but there's also information on the Zen 4-based EPYC processors as well, right down to the layout of the compute tiles.
Intel has now formally introduced its new ARC series graphics cards -- at least by naming scheme only, with a nice marketing trick teasing the ARC GPU with 1000 drones all synched together. Check it out:
The design of the new Intel ARC series graphics cards were teased by drones, and looks virtually identical to the leaked photos that Tom from Moore's Law is Dead shared months ago, back in April 2021. It appears he was right, again. The upcoming Intel ARC series GPU had 9 blades in the leak... and 9 blades shown by the drones as well.
Intel's new Core i9-12900K processor is still in the oven and months away, but it has appeared in the CPU black market -- and now on the BaseMark benchmark database.
The new Alder Lake-S flagship processor has 12 cores and a CPU clock of 3.2GHz on the BaseMark database, but we've been hearing it'll be a 16-core, 24-thread CPU. This will see 8 high-performance cores (with HT) and 8 high-efficiency cores for 24 threads in total on the Core i9-12900K processor.
Intel's Core i9-12900K processor was reportedly run on an "Acer Z69H6-AM" which going by the Z69 part of the product name could mean a new Acer Z690 motherboard which would be capable of taking a new 12th Gen Core "Alder Lake" CPU.
We're hearing new rumors that Intel has secured most of TSMC's upcoming 3nm production capacity, underscoring rumors from all the way back in January 2021.
Now we have Chinese publication UDN reporting that mass production of Intel's new CPUs on TSMC's new 3nm node will begin in Q2 2022, with production capacity expected to hit 4000 wafers in May 2022, and in the months after that, it will flow out 10,000 wafers per month.
The new report says that Intel will be using TSMC's new 3nm node for 3 x GPUs, and 1 x GPU -- we should expect the 3nm CPUs to be next-gen Xeon processors for next-gen servers and data centers -- not the upcoming Raptor Lake CPUs that will be launching in the second half of 2022.
AMD's new Ryzen 3 5300G processor isn't the one you'd think of when it comes to gaming your days and nights away on, but for an overclocker... well, holy sh*t -- the APU has just been OC'd all the way up to an insane record-breaking 5.5GHz.
The Ryzen 3 5300G is a quad-core APU that you can buy for around $110-$130 in the US, with a Vega-based integrated GPU for gamers that don't have deeper pockets for a higher-end CPU and GPU combination. Well, during an overclocking marathon by Safedick, who used an ASUS ROG Crosshair VIII Extreme motherboard and 16 GB of G.Skill Ripjaws V DDR4-2066 (CL14-14-14-28-1T) memory.
With that combo, the AMD Ryzen 3 5300G processor was cranked up to 5.5GHz under LN2 cooling, so don't be thinking you'd be pushing that OC under a normal air cooler or AIO liquid cooler.
We are receiving drops and drops of information on Google's next-gen ultra-high-end Pixel 6 and Pixel 6 Pro smartphones, where they will be manufactured using Samsung's next-gen 5nm process node.
Google is using a custom in-house Tensor chip inside of its next-gen Pixel 6 and Pixel 6 Pro smartphones, with new rumors suggesting that Samsung will be making the chip using its new 5nm process technology. The news is coming from sources that go against Nikkei, that reported TSMC was going to be building the Tensor chips for Google.
The new Tensor core is going to built on Samsung's new 5nm process according to the latest reports, but of course -- Google and Samsung are refusing to comment. Google has been silent on the inner happenings of its new Pixel 6 and Pixel 6 Pro smartphones as it is, so I'm sure they're going to be even more hush-hush when it comes to the collaboration with Samsung on making its new Tensor chips on their new 5nm node.