Hardware

The patent was filed 18 months ago, and there’s no further indication of ongoing development

Sharon AI and Firmus are the first partners, with a combined total of up to 210,000 GPUs.

Alleged illicit GPU movements lead to seizure of $42 million house

Samsung Foundry has reportedly resumed 1.4nm chip development after improving its 2nm yields, with mass production targeted for 2029. - SamMobile

Ampere rides again for almost the same price as RTX 5060s

Samsung, SK Hynix and Micron have been accused of fixing RAM prices and supply, thereby exacerbating the memory crisis, in US anti-trust lawsuit.

Even with lots of RAM, GPU, and fast disks, you probably don’t want it

It uses two wafers instead of one, along with ultra-thin dielectric bonding

CXMT IPO, SKHynix, Micron, Samsung Competition, Process Node Deficit, China HBM, Wafer Adds, Memory LTAs

The company is in discussions with manufacturers of smartphones, personal computers, and cars about an aspect of its new data center chip technology.

Pro, Max, and Ultra tiers move to the M7 line as Apple chases memory bandwidth for on-device AI.

PLUS: Indian telco ponders broadband satellites; Samsung goes all-in on OpenAI; Vietnam centrally plans ten tech giants; and more!

Dimming the lights down on the TV room may not be a necessity for much longer.

Running AI models on x86 CPUs is becoming easier and faster

Following up from the HD 4850 video — this time a reference HD 4870 for $11. Tested on 6 games on the Core 2 Duo rig.

Google’s new smart speaker is more about Gemini than audio quality.

The new "Fractal" operating system kernel, developed at MIT, gives researchers a clearer look at how computer chips work internally. It was used to study the branch predictors inside Apple's M1 processor, where it revealed a potential vulnerability to major speculative attacks.

Remembering the ups and downs of the Intel Mac era as it finally winds down.

AMD’s David McAfee, VP and general manager of Radeon and Ryzen, says that “a whole body of engineering work” went into the re-release, as the original bonding process TSMC used for the Ryzen 7 5800X3D was no longer available. […]

A surprisingly large revision of 18A.

But NAND supply for client devices may not improve for years.

Demand for AI systems plus the shortage of DRAM and NAND are shaping the global market

Linux-based phone still ‘runs 99% of Android apps’ so you can do more with it, if you wish.

But for what?

Following up from my HD 3850 video — this time a Gigabyte HD 4850 with Zalman aftermarket cooler for $10. Cleaned it, repasted it and tested on 6 games on a Core 2 Duo rig.

Two years in the making

Chinese cloud firms are reportedly testing more than 300 Vera servers overseas.

We’re moving as fast as we can, says SK Group chair

Desperate times call for old silicon

The Mobi Fold is an $80 Bluetooth mouse with a silicone-wrapped hinge.

The AMD Ryzen 9 9950X3D2 Dual Edition sets a new standard for desktop performance, featuring 3D V-Cache on both chiplets for unprecedented gaming and productivity gains. With 16 cores paired with 192 MB of L3 cache, this processor delivers exceptional performance without the complexity of asymmetric caching. Yes, it dominates gaming, too.

Technology may help circumvent US export restrictions

Reports from industry sources suggest NVIDIA has revived its GeForce RTX 50 SUPER graphics card program, potentially paving the way for a memory-enhanced refresh of several Blackwell-based gaming GPUs.

Passing the $650 billion record Novo Nordisk set in June 2024.

Back to the (stone) DDR4 age.

Nvidia has stepped into the processor market with its RTX Spark, but at first glance, it’s clearly behind Apple Silicon by a considerable margin.

Nvidia’s new chips will power laptop workstations and mini desktop PCs at first.

Chipzilla’s first 2nm-class Xeon is finally here bristling with 288 cores

The latest twist in paying humans to wear head cameras for robot training data.

Intel has lined up partners for the range, including Acer and OneXPlayer.

The paper tiger that’s flying off the hardware shelves.

Nvidia’s x86 CPU success rests on vertical integration.

This is to challenge the creativity of the computer lover.

It could be one way to make your old PC play nicely with a high-end GPU.

Revenue is at record highs, but even that isn’t enough to fund the inventory these companies need.