NVIDIA & AMD Planning To Launch Arm-Based CPUs For PCs, To Tackle Intel & Apple By 2025
The report comes from Reuters who have cited their own sources on the matter, stating that NVIDIA has internally started prep work on its next-gen CPUs for the client PC segment which will target Intel and run on the Arm architecture. The company not only wants to take on Intel but also tackle Apple in the client segment
But it isn't just NVIDIA who will be making Arm-based CPUs, AMD is also reported to be going big with Arm in its future CPUs. The company is also said to introduce Arm-based CPUs which will sit alongside its traditional x86 CPUs serving the client consumer PC space under the Ryzen banner.
Arm-based CPUs should run the Windows operating systems.
AMD has worked on Arm CPUs in the past -- but, stopped work to focus on x86 CPUs.
AMD CEO Says EPYC Turin “Zen 5” CPUs Deliver Significant Performance & Efficiency Gains, MI300 To Drive Over $2B Revenue In 2024, Talks Arm PC Competition
During their Q3 2023 earnings call, AMD CEO, Dr. Lisa Su, talked about the upcoming EPYC "Turin" CPU & Instinct "MI300" GPU portfolio for Data Centers/AI while also commenting on the surging competition from Arm-based chips in the PC segment.
This was about "commercial"/"professional" CPU/APU usage. AMD says new products are in advanced development and "on the way" in 2024.
AI Company Plans to Run Clusters of 10,000 Nvidia H100 [AI] GPUs in International Waters
Del Complex hopes floating its [Artificial Intelligence] computer clusters in the middle of the ocean will allow it a level of autonomy unlikely to be found on land.
Government entities around the globe are scrambling to implement AI laws and best practices as the technology grows in both strength and scope. The companies that make and use those AI systems, however, aren’t too thrilled about how government oversight might impact business. Del Complex is one of them—and it has a unique plan to circumvent whatever regulatory challenges are coming.
The United States-based “alternate reality” company, which seems to be a catch-all for AI, clean energy, neural prosthetics, robotics, and “technological resilience” research, has decided to float its own AI-dedicated computer clusters in international waters.
Nvidia AI GPUs lead on the AI front. AMD has developed an AI GPU to compete.
This article highlights that AI usage is limited in each Country by Governmental control. So, AI developers are developing "workarounds" -- like not being in a Country.
Unlike the competing solutions, AMD's FSR technology works with a range of hardware including GPUs from the competition. This hardware-agnostic design can also serve the mobile space well and it looks like Samsung & Qualcomm have partnered up with AMD to utilize the tech for their upcoming flagship smartphones.
The use of AMD FSR within smartphones will definitely be a major development for Arm & Android-based devices. Expect more information in the coming months as these products are introduced to the consumer segment.
FWIW: I recently purchased Samsung's current generation flagship smartphone.
It does not have FSR. However, I won't be buying a new one anytime soon.
Intel Kills Off Its Chip-Freezing Cryo Cooling Technology
Intel discontinued the company's Cryo Cooling Technology on July 1, 2023, with the 13th Generation Raptor Lake parts being the last supported chips.
In a nutshell, this cooling tech brings the CPU down to 0 degrees Celsius to boost overall performance.
The thing with Intel's Cryo Cooling Technology is that it wasn't just about the hardware. Intel had envisioned it to be a small ecosystem comprised of software, firmware, and hardware working in harmony. The chipmaker's unexpected move to discontinue Cryo Cooling Technology puts its partners in a difficult predicament.
Intel a Candidate for Billions More in CHIPS Act Funding for Defense Work
Intel ... is apparently involved in a secret 'secure enclave' project designed to ensure that the U.S. military has access to chips produced in America, reports the Wall Street Journal.
Intel is currently building four new fabs in the U.S.: .. near Chandler, Arizona, that are due to come online in 2024 as well as two .. near Columbus, Ohio, that are set to start mass production in 2025. Both vast projects require investments of $100 billion or more, so it is not surprising why Intel needs government support.
[However, there is concern about spending "all" the money only with Intel] "We are concerned that the [Department of Defence] is considering sole sourcing an award to one company to build a secure enclave at a cost that is far greater than [a typical approach to secure chips for defence applications]," a letter obtained by The Wall Street Journal and signed by multiple American politicians reads. "Doing so would limit funding for other projects that would create a diversified domestic supplier base of semiconductors critical to the defense industrial base."
Secure production in the USA is good because China could halt production for the USA in Asia at any time.
However, long term, Some are concerned that it may be better/safer/fairer to multi-source/diversify in the USA with several manufacturers -- not just Intel alone.
The latest rumor on AMD's highly anticipated AM5 Desktop APUs comes from HKEPC who state that AMD will be launching its Ryzen 8000G APU family soon.
Their information comes from a Taiwanese-based motherboard manufacturer who has received the first ES Ryzen "AM5" APUs for testing and based on the information, the chips will be marketed as Ryzen 8000G instead of Ryzen 7000G which was the previously rumored branding.
Had trouble with my high end GIGABYTE X670E AORUS Master MB. So, got new Gigabyte X670 AORUS Elite AX -- one notch lower and much less cost. Runs the same.
New setup includes a new case and new cooling. The signature is updated.
Meantime, I'll consider sending the "old" GIGABYTE X670E AORUS Master MB for repair or replacement -- for the SECOND time. They fixed it the last time. It was the first MB that ever gave me a serious issue (I've had "dozens" over the last 23 years).
AMD To Utilize Samsung 4nm & TSMC 3nm Nodes For Next-Gen Chips, Zen 5C Possibly Codenamed Prometheus
We know that AMD is going to use a mix of 4nm and 3nm process nodes for its Zen 5 core architecture but the company has so far relied on TSMC for production.
The AMD Zen 5 and Zen 5C core architectures will be a major deal in 2024-2025. They are going to power a range of families including Strix Point (Ryzen Laptops), Granite Ridge (Ryzen Desktop), and Turin (EPYC Server).
The nearby launch of Ryzen 8000G series APUs get more confirmation.
Quote
Quote
GIGABYTE TECHNOLOGY Co. Ltd, a leading manufacturer of motherboards, graphics cards, and hardware solutions, released the latest AGESA 1.1.0.0 beta bios for AM5 next gen APU support on X670, B650, A620 motherboards. The forthcoming AM5 next gen APU will be launched at the end of January 2024.
According to recent reports, the next-gen AMD Ryzen APU lineup for AM5 desktops will launch under the Ryzen 8000 branding and will include a total of four SKUs featuring the Zen 4 & Zen 4C core configurations. This will be the first desktop APU lineup that supports DDR5 memory and PCIe Gen 5.0 connectivity. Pairing them up with the A620 and B650 series of motherboards will make the most sense since these APUs are targeted at budget and mainstream audiences.
Hundreds of RTX 4090s With Melted Power Connectors Repaired Every Month, Says Technician
The saga continues over a year after the first problems with GeForce RTX 4090's melted 12VHPWR connectors were reported.
A NorthridgeFix repairman claims he must fix about a hundred GeForce RTX 4090 graphics cards with failed connectors every month. He keeps saying that the connector failures are not user error, but issues with the connector design.
Back in November 2022, Nvidia said that it was aware of 50 cases of melted GeForce RTX 4090 power connectors and that the standard issue was that they were not fully plugged into the graphics card, which essentially meant that it was a user error, not a failure of the connector.
Quote
"So now it is safe to say the [GeForce RTX] 4090 melted connector is not a user problem," the repairman said. "We discussed this many times in the past, but I want to mention one last time that this is not a user error."
AMD Announces Ryzen Embedded 7000 Series Processors Powered by Zen 4
AMD today announced at Smart Production Solutions 2023 the AMD Ryzen Embedded 7000 Series processor family, optimized for the high-performance requirements of industrial markets.
The Ryzen Embedded 7000 Series processor is the first embedded processor to use next-generation 5 nm technology with a 7-year manufacturing availability commitment. The new embedded processor integrates AMD Radeon RDNA 2 graphics that eliminates the need for a discrete GPU for industrial applications. And because embedded applications require additional operating system software options, Ryzen Embedded 7000 Series processors include support for both Windows Server and Linux Ubuntu, on top of Windows 10 and Windows 11. Ryzen Embedded 7000 Series processors also include up to 12 high-performance "Zen 4" CPU cores, which combined with its integrated features and wide operating system choices, offers unparalleled ease of integration for system designers.
These are not aimed at desktop PCs. Rather, they will be used to make industrial and scientific equipment. They will use the AM5 socket -- like a PC -- to simplify integration into equipment.
NVIDIA GeForce RTX 50 Flagship Gaming GPU Rumored To Feature GDDR7 Memory & 384-bit Bus
The NVIDIA GeForce RTX 50 series is expected to utilize the Blackwell Gaming GPU architecture though a bit different from the one coming to HPC and AI customers next year. Currently, there have been roadmaps that point out to a 2025 launch but some rumors suggest that late 2024 is when we might get to see the latest graphics lineup.
Although the exact memory configurations aren't mentioned for the NVIDIA GeForce RTX 50 Flagship GPU, what is mentioned is that the card will feature a 384-bit bus interface so it seems like 512-bit may no longer be on the planning board or it is just meant for an even higher-end variant.
AMD is rumored to NOT have a high-end, expensive card in its next desktop generation. Rather, the "best" card will be a mid priced card that plays the games but does not provide "over the top" FPS. They say that's a conscious business decision -- not because they can't make one.
AMD is said to be putting the effort into advanced industrial cards. That's because there will not be enough chips to do both desktop PCs and top industry GPUs (that cost a lot to buy; and, thus, make AMD a lot of money).
AMD Reportedly Taps Samsung Foundry for 4nm Zen 5c Designs
The company is diversifying its chip sourcing, according to a new report.
AMD has traditionally used TSMC to craft all of its most advanced silicon designs, but now that it's working on its next-gen Zen 5 architecture, it looks to be switching things up a bit. A new report from Taiwan states AMD has enlisted Samsung Foundry to produce Zen 5c products for its next-generation platform, which marks a huge shift for the company. It will still use TSMC as well, but it appears to want to diversify its sourcing for Zen 5 instead of relying solely on TSMC as it has done with Zen 4 and the previous generations.
{AMD will] utilize Samsung's 4nm process for "basic" versions of its future chips while allowing TSMC to make more advanced variations on its cutting-edge 3nm node.
NVIDIA GeForce RTX 4090 GPU Prices Hit Close To $2000 In The US, Cheapest Model Now At $1900 US
Back in October, the lowest price for the card reached $1700 US and now that sits at $1900 US, a $200 US increase in just a few weeks. Furthermore, the $1900 US pricing is limited to Black Friday promos which means that once those deals are over, you will see the card listed for even higher prices. The average price at the moment for the card has reached $2000 US.
If we do a comparison between the N5 (4N Optimized 5nm variant for NVIDIA), the latest 3nm process technology would offer a 25-30% reduction in power, 10-15% additional performance per transistor, 42% reduction in area and a density increase of 1.7x. Those are quite substantial gains over the current 5nm node which offers some amazing performance per watt (efficiency) in the existing GeForce RTX 40 "Ada Lovelace" lineup so we can only expect NVIDIA's efficiency lead over its competition to further increase.
AI (Artificial Intelligence) is in the process of rapid growth. Gaming GPUs can be used (although special AI GPUs are made for the purpose).
Quote
AMD Radeon RX 7900 XT Is The Latest RDNA 3 GPU To Land Support For ROCm 5.7, Accelerated AI With PyTorch
Today's announcement from AMD builds upon its promise to deliver support for ROCm and other machine learning and development features to its consumer-centric products within the RDNA 3 GPU family. The red team had already introduced support for the Radeon RX 7900 XTX & Radeon PRO W7900 GPUs last month and now, it is adding the Radeon RX 7900 XT too which covers the top echelon of RDNA 3 products.
FWIW: I recently mentioned "Had [un-diagnosable] trouble with my high end GIGABYTE X670E AORUS Master MB. So, got new Gigabyte X670 AORUS Elite AX"
Meantime, just to see if I could somehow fix things, I did a "from scratch" rebuild using my "nominally defective" GIGABYTE X670E AORUS Master MB. It included some changes from the original build.
So far, its been working a couple days. So, I've changed the signature line to show the rebuilt system. We'll see.
AMD AM4 To Live On In 2024 As Red Team Preps More 3D V-Cache CPUs & APUs For Desktop Platform
AMD is likely preparing at least five brand new SKUs which will include the Ryzen 7 5700X3D, Ryzen 5 5600X3D, Ryzen 5 5600GT, Ryzen 5 5500GT and the Ryzen 7 5700.
These new SKUs will ensure that mainstream audiences stick with AM4 for a good while and enjoy the gaming performance uplift offered by the addition of 3D V-Cache tech. Besides the SKUs, it may look like AMD is planning to replace its existing Ryzen 5 5600G APU with the Ryzen 5 5600GT & the Ryzen 5 5500GT. The "GT" naming is definitely new in the Ryzen APU family and it is likely that the two chips will offer a slight upgrade.
AMD Ryzen 8000G AM5 Desktop APU Specifications Leak, Up To 2.5x Faster Than Ryzen 5000G In Gaming Benchmarks
The new APUs make use of two key technologies, first is the Zen 4 core architecture which will power the CPU side of things and the second is the RDNA 3 core architecture which will power the graphics. The new family will be supported on AM5 motherboards using the latest BIOS which has already been rolled out across multiple 600-series products
the AMD Ryzen 8000G APUs will be available in both PRO and Non-PRO flavors which share identical specifications. The flagship SKU is going to be the Ryzen 7 8700G which would offer 8 cores, 16 threads, a base clock of 4.2 GHz, and a boost clock of 5.1 GHz. There's also the Ryzen 5 8600G APU which features 6 cores, 12 threads, a base clock of 4.35 GHz & a boost clock of 5.0 GHz. Both of these APUs will be based on the Phoenix 1 die and feature support for OC.
The Phoenix 2 SKUs will include the Ryzen 5 8500G with 6 cores and 12 threads and the Ryzen 3 8300G with 4 cores and 8 threads. The 8500G operates at a base clock of 3.55 GHz and a boost clock of 5.0 GHz while the 8300G operates at a base clock of 3.45 GHz and a boost clock of 4.90 GHz.
