hardware
Analog In-memory Training on General Non-ideal Resistive Elements: The Impact of Response Functions
As the economic and environmental costs of training and deploying large vision or language models increase dramatically, analog in-memory computing (AIMC) emerges as a promising energy-efficient solution. However, the training perspective, especially its training dynamics, is underexplored. In AIMC hardware, the trainable weights are represented by the conductance of resistive elements and updated using consecutive electrical pulses. While the conductance changes by a constant in response to each pulse, in reality, the change is scaled by asymmetric and non-linear response functions, leading to a non-ideal training dynamics. This paper provides a theoretical foundation for gradient-based training on AIMC hardware with nonideal response functions.
PC building's weird new reality: Your favorite old parts are back on the menu
PCWorld reports that rising RAM and storage prices are driving hardware vendors to re-release older components like AMD's Ryzen 7 5800X3D and GeForce RTX 30-series cards. This trend matters for PC builders seeking affordable alternatives as memory shortages and cost increases make newer hardware less accessible. The shift encourages enthusiasts to find creative solutions with existing components and embrace the joy of tinkering despite market challenges. Your weekly edition of The Full Nerd has arrived, and there's a new face on the team: mine!
You Can Finally Buy Snap's New AR Specs--for 2,150
You Can Finally Buy Snap's New AR Specs--for $2,195 Snap CEO Evan Spiegel lays out the company's vision for its augmented-reality smart glasses, arriving later this year. Snap--maker of the popular social app Snapchat--has a new pair of augmented-reality smart glasses called Specs. Snap CEO Evan Spiegel revealed the new glasses at an event during the Augmented World Expo (AWE) tech conference in Long Beach, California. As Snap frames it, this isn't a prototype or developer device--it's the first actual consumer version of the Specs AR glasses, unlike the previous generation exclusively sold to developers and creators. Snap says it expects the devices to ship this fall in the US, UK, and France.
SteamOS is coming for Intel handhelds -- if Intel can keep up
PCWorld reports that Valve's SteamOS is now available in beta for Intel-based handhelds, starting with the MSI Claw, potentially challenging Microsoft's Windows dominance in PC gaming. Intel's new Arc G3 processors are debuting in handhelds from MSI, Acer, and OneXPlayer, aiming to compete with AMD in the portable gaming market. Early benchmarks show SteamOS performing slightly behind Windows 11 on Intel devices, but this expansion could establish SteamOS as the unofficial standard for PC gaming handhelds. The Steam Deck didn't invent the handheld gaming form factor, or even debut it for PC hardware, but it's certainly the iPhone equivalent for this particular moment. And the vast, vast majority of the Steam Deck-inspired market has been underpinned by AMD's integrated chips.
Analog Foundation Models
Analog in-memory computing (AIMC) is a promising compute paradigm to improve speed and power efficiency of neural network inference beyond the limits of conventional von Neumann-based architectures. However, AIMC introduces fundamental challenges such as noisy computations and strict constraints on input and output quantization. Because of these constraints and imprecisions, off-the-shelf LLMs are not able to achieve 4-bit-level performance when deployed on AIMC-based hardware. While researchers previously investigated recovering this accuracy gap on small, mostly vision-based models, a generic method applicable to LLMs pre-trained on trillions of tokens does not yet exist. In this work, we introduce a general and scalable method to robustly adapt LLMs for execution on noisy, low-precision analog hardware. Our approach enables state-of-the-art models -- including Phi-3-mini-4k-instruct and Llama-3.2-1B-Instruct
Nvidia's N1X could show us the future of PCs--and the bill that comes with it
PCWorld anticipates Nvidia's N1X launch at Computex, featuring an Arm-based APU with 20 CPU cores and Blackwell graphics that could match RTX 5060 laptop performance. The article highlights growing concerns about PC hardware affordability, with examples like Steam Deck price increases suggesting higher costs may become the norm. This trend matters for consumers as powerful new hardware from Nvidia, AMD, and Intel may deliver impressive performance but potentially at premium prices that limit accessibility. The PC industry is once again on the brink of a pivotal moment in history--or so appears to be the case, given the rumors about Computex next week. In particular, the internet anticipates the launch of Nvidia's N1X, an Arm-based APU expected to marry ferocious CPU performance with equally knockout GPU chops.
Your HDR monitor might be lying to you
Many monitors advertise HDR support but lack essential hardware like sufficient brightness (1000 nits recommended) and proper local dimming to deliver true HDR experiences. PCWorld highlights that even high-end models like the Alienware AW3225QF may only achieve 250 nits in full-screen HDR despite 1000-nit peak claims. Consumers should prioritize raw specifications over marketing terms, as technologies like OLED or Mini-LED with strong contrast ratios are crucial for authentic HDR performance. Imagine you just bought a brand-new monitor that prominently advertises its HDR capabilities. You hop over to YouTube and stream the first "4K HDR" video you find but it looks washed out or barely any different from the non-HDR display you had before.
I'll never buy another laptop without Windows Hello
When you purchase through links in our articles, we may earn a small commission. I'll never buy another laptop without Windows Hello It's so important, I even go out of my way to add it to my desktop PCs! I review a lot of laptops here at PCWorld, and I'm always surprised when I unbox one without biometric hardware. Fingerprint readers and infrared (IR) cameras for facial recognition are both great--and some laptops even have both! But then you have those that skip the biometric hardware entirely, which is a huge loss these days.