GPU and CPU news and discussions

[wjfox]

Microsoft says its newest AI chip Maia 200 is 3 times more powerful than Google's TPU and Amazon's Trainium processor

By Roland Moore-Colyer
published 60 minutes ago

The Maia 200 AI chip is described as an inference powerhouse — meaning it could lead AI models to apply their knowledge to real-world situations much faster and more efficiently.

Microsoft has revealed its new Maia 200 accelerator chip for artificial intelligence (AI) that is three times more powerful than hardware from rivals like Google and Amazon, company representatives say.

This newest chip will be used in AI inference rather than training, powering systems and agents used to make predictions, provide answers to queries and generate outputs based on new data that's fed to them.

Maia 200 is already being deployed in Microsoft's U.S. central data center region, with the company set to use the chips to generate synthetic data and in reinforcement training to improve next-generation large language models (LLMs). The AI accelerator will also be used to power Microsoft Foundry and 365 Copilot AI, and be part of the infrastructure that the company can provide through its Azure cloud platform.

The new chip delivers performance of more than 10 petaflops (1015 floating point operations per second), Scott Guthrie, cloud and AI executive vice president at Microsoft, said in a blog post. This is a measure of performance in supercomputing, where the most powerful supercomputers in the world can reach more than 1,000 petaflops of power.

https://www.livescience.com/technology/ ... -processor



Credit: Microsoft

[firestar464]

This means Azure OpenAI users, such as scientists, developers and corporations, could see better throughput and speeds when developing AI applications and using the likes of GPT-4 in their operations.

Bro, someone is really late to the party

[firestar464]

[wjfox]

[weatheriscool]

AMD Breaks GPU Frequency Record With 4.77GHz 9060 XT
They beat the previous record by over 500MHz!
By Jon Martindale February 20, 2026

AMD recently teamed up with overclockers Bill Alverson and Allen 'Splave' Golibersuch to go after the GPU frequency record, and they've shared their success in a recent video. In it, they pushed an RX 9060 XT to its limits with liquid nitrogen and enhanced power delivery, hitting an utterly insane 4.769GHz. That's not just a new record; it demolishes the old one and puts AMD within spitting distance of the 5GHz threshold that would have seemed impossible just a generation or two ago.

Extreme GPU overclocking has made a big comeback with this generation of graphics cards. That's partly because neither AMD nor Nvidia is releasing new models while they try to fill their overbought orders for AI data centers, and partly because this latest batch of GPUs is power-hungry monsters. Although the 9060 XT is an entry-level GPU targeting 1080p and 1440p play, it can still handle a lot of juice if you keep it cool enough.

https://www.extremetech.com/computing/a ... hz-9060-xt

[weatheriscool]

Intel Announces 290HX and 270HX High-End Arrow Lake Refresh CPUs
A near-1GHz boost to the die-to-die interconnect may be behind the performance boost.
By Jon Martindale March 17, 2026
https://www.extremetech.com/computing/i ... fresh-cpus

Following the announcement of its refreshed Arrow Lake desktop processors last week, Intel has debuted two new high-end mobile CPUs in the same vein.

The Core Ultra 9 290HX is a 24-core CPU with a maximum boost clock of 5.5GHz. The Core Ultra 7 270HX has 20 cores and a max boost clock of 5.3GHz, per VideoCardz. Although both chips only enjoy a 100MHz bump on the boost clock over their predecessors, they do get a 900MHz increase on the die-to-die interconnect, which could explain the near-10% gaming performance bump Intel is claiming.

Arrow Lake processors aren't best known for their gaming ability, even though they are perfectly serviceable at it. Intel is talking up these new chips' capabilities, however, claiming an average of 8% higher frame rates compared to the 285HX and 265HX, respectively, across 32 different games at 1080p. It also claims a 7% higher single-thread performance in Cinebench 2026 and that, following these innovations, these CPUs are now up to 62% faster at gaming than Intel's venerable 12900HX from a few generations prior.

[weatheriscool]

[weatheriscool]

[weatheriscool]

Intel's Low-Cost Wildcat CPUs Leak, Revealing Core, Clock, and TDP Details
These could be some of the most impressive little CPUs ever.
By Jon Martindale April 7, 2026

Intel's rumored Wildcat CPUs have leaked again, this time in a more impressive and detailed fashion, showcasing some intriguing low-power, low-core-count designs that could be great for embedded and edge-case settings. These new CPUs will come in a few different flavors, each sporting six cores across a pair of architectures, with a maximum TDP of just 15W, Tom's Hardware reports.

This latest leak comes from an OEM called Advantech, which showed off a new product featuring a trio of CPU options, which were really unannounced Wildcat chips. They include the Core 7 350, Core 5 320, and Core 3 305, each comprising two performance cores and four low-power-efficiency cores—no standard E-cores in sight with these processors.

They sport the same 1.5GHz base clock across the board, though the higher-end Core i7 chip can reach 4.8GHz, while the others can achieve 4.6GHz and 4.4GHz, respectively. All chips support the same DDR5-6400 memory and have a capacity of up to 64GB. The only other difference between the chips is in graphics: The Core 7 and Core 5 models enjoy a 32 Execution Unit GPU, while the Core 3 has just 16 EUs. The Core 7 has a 100 MHz higher GPU boost clock, too.

https://www.extremetech.com/computing/i ... dp-details

[weatheriscool]

Intel's Nova Lake Lineup Is Gunning for AMD's Gaming Crown
With up to 52 cores, 12 Xe 3 GPU cores, a big cache, and support for faster memory, Nova Lake could be exceptionally quick.
By Jon Martindale April 14, 2026
https://www.extremetech.com/computing/i ... ming-crown

Intel hasn't competed for flagship gaming performance for a couple of generations now, but its next-generation Nova Lake chips could be the fastest ever when they drop later this year. Based on VideoCardz's preliminary look at the product stack, it seems clear Intel is going after AMD's gaming dominance in a big way with some seriously powerful processors.

The entire Nova Lake lineup is moving to a new socket, LGA 1954, with more pins than ever before. That's in aid of increased memory and PCIe 5 support, but it's just the start of what's changing with these next-gen chips. There are new architectures everywhere, including the new Coyote Cove P-cores, Arctic Wolf E-cores, a next-generation Intel NPU 6 neural processor, and Xe3p graphics.

The top chip will have 52 cores in a dual-die configuration, with 16 P cores, 32 E cores, and 4 low-power cores, and a big-cache chip attached to deliver AMD X3D-style gaming performance. There will also be a 28-core variant, a single-die 28-core chip, and 16-core and 8-core options. Although all of them will sport Xe3p graphics, it'll mostly be just two cores there. There's also a rumored variant with 16 CPU cores and 12 Xe3p GPU cores, which could make it a very fast entry-level gaming CPU.

[weatheriscool]

[weatheriscool]

[weatheriscool]

[weatheriscool]

[weatheriscool]

[weatheriscool]