Digital Design & Computer Architecture - Lecture 23: Memory Hierarchy and Caches (Spring 2022)
The lecture reviewed fundamentals of memory organization and the design of memory hierarchies and caches, emphasizing why SRAM is used for on-chip caches while DRAM serves as main memory due to differing fabrication and capacitor requirements. It surveyed memory technologies — DRAM, SRAM, flash, phase-change (PCM) and emerging resistive memories — highlighting trade-offs such as latency versus density, non-volatility, and endurance constraints. The instructor noted how technologies like flash and PCM can disrupt traditional hierarchies by blurring the line between working memory and storage, and described the complexity of flash-based storage controllers and their internal memory hierarchies. Practical manufacturing and system-design considerations were stressed as key reasons for current architectural choices and potential future shifts.
Converting an ASUS GTX 950 Strix to Run on only PCI-E Slot Power.
YouTuber BuildZoid documents a hardware mod converting an ASUS GTX 950 Strix to run solely from the PCIe slot by raising the shunt resistor for the auxiliary 6‑pin feed and rewiring the card so the 6‑pin, when plugged in, bypasses...
How NOT to Repair an iPhone 17 Pro
The video is a tongue‑in‑cheek tutorial on what not to do when repairing an iPhone 17 Pro, focusing on the delicate process of screen removal and the pitfalls that can instantly ruin the device. The presenter walks through each misstep: stripping the pentalobe...
Picking up a Used Ryzen 5 7600
The video documents a vlog‑style purchase of a used AMD Ryzen 5 7600, positioning it as a budget alternative to the newer X3D‑enabled 9800 and 9850 models. After acquiring the six‑core, twelve‑thread chip at roughly a third of the price of its...
Nova Lake Might Not Suck - Tech News Feb 15
Paul’s weekly tech roundup zeroes in on three headline‑grabbing developments: Intel’s upcoming Nova Lake desktop processors, MSI’s ultra‑premium RTX 5090 Lightning graphics card, and Discord’s controversial age‑verification rollout. The episode also touches on the lingering RAM shortage and a tongue‑in‑cheek “ZAM”...
Galaxy Fold 8 WIDE Revealed??
A leaked animation from Android Authority and Assemble Debug suggests Samsung is preparing a new “Fold 8 Wide” device, slated to debut alongside the Flip 8 and Fold 8 later this summer. The animation reveals a 4:3 inner display, matching iPad dimensions and ideal for...
Nvidia Retail Prices Might Be Over
The video reports that Nvidia’s Open Price Program (OPP), a rebate scheme that kept entry‑level graphics cards near the manufacturer’s suggested retail price, is being phased out. Under OPP, partners such as Asus, Gigabyte and MSI received direct or delayed rebates...
Best DTF Printer to Start a T-Shirt Business in 2026
Starting a t‑shirt printing venture in 2026 hinges on selecting the right direct‑to‑film (DTF) printer. For newcomers, the R2 Pro offers a practical entry point, handling roughly 200 transfers per day without straining workflow. As order volumes grow, the industrial‑grade...
PCPer Podcast 856: ASRock BIOS Fix, 700W Intel CPU? HP Rents Gaming Laptops, Homelab Bling, and MORE
The PC Perspective episode 856 covered a wide range of hardware and policy news, from a critical BIOS update for ASRock AM5 boards to Intel’s looming 700‑watt power draw on its Nova Lake desktop CPUs. The hosts also dissected a...
$5090 RTX 5090 Lightning Unboxing
The video showcases the unboxing of Nvidia’s RTX 5090 Lightning, positioned as the fastest gaming graphics card on the market. The presenter emphasizes its headline‑grabbing performance claim and immediately highlights the card’s substantial cooling solution. Benchmarks cited in the clip suggest frame‑rate...
Realme P4 Power Battery Test: Our New Battery Life Champion
The Realme P4 Power’s headline is its colossal 10,000 mAh battery, a rarity in the mid‑range segment. In a dedicated battery‑life suite the phone posted an active‑use score exceeding 25 hours, translating to roughly two full days of screen‑on time, despite...
Memory Centric Computing - Sonova Workshop, 24.10.2022
In a workshop on memory-centric computing, the speaker argued that modern computing is bottlenecked by data movement rather than raw compute, urging co-design of hardware and software to keep memory and compute tightly coupled. He highlighted neural networks and genome...
Reviewing the OFFICIAL Windows Gaming PC
The video reviews Microsoft’s official Windows gaming PC guide, outlining three pre‑defined build tiers—entry‑level at roughly $600, mid‑range around $900, and high‑end near $1,500. It positions the guide as a practical roadmap for DIY enthusiasts seeking a Windows‑optimized gaming rig. Key...
UniFi UNAS Drive 4.0 (and WHY SYNOLOGY SHOULD BE WORRIED)!
Ubiquiti has released UniFi Drive 4.0 (currently in staged release candidate) for its UNAS lineup, adding native iPhone photo backup, richer drive analytics and integrated logging, and one-click SMB access via Identity Endpoint including mobile access. The headline feature is...
Testing Ryzen DDR5 Refresh Modes with Dual Rank Hynix 24Gb M-Die
The video examines how the new DDR5 refresh‑mode options on AMD’s AM5 platform affect performance with a dual‑rank 24 Gb Hynix M‑die kit (Corsair 2 × 48 GB 6000 CL30). Using an ASUS Crosshair X870 Hero with BIOS 20004 (AGSA 1.3.0.0.0), Buildzoid compares the standard "normal" mode, the...
1440p WOLED vs QD-OLED in 2026 - Everything You Need to Know
The video dissects the 2026 landscape of 1440p gaming OLED monitors, contrasting Samsung’s QD‑OLED line with LG’s third‑generation tandem W‑OLED panels. It maps eight panel variants—four QD‑OLED (240‑500 Hz) and four W‑OLED (240‑540 Hz)—and explains how refresh rate, sub‑pixel architecture, and surface...
Digital Design & Computer Arch. - Lecture 22: Memory Organization & Memory Technology (Spring 2022)
In this lecture the professor shifts focus from processors to memory, arguing that memory and storage are the dominant bottlenecks in modern computing. He frames the discussion with Amdahl’s law to show why accelerating computation alone yields limited system speedups...
Multi-Frame Generation on Intel Arc
Intel’s latest Arc drivers add multi-frame generation (up to 3x power) to Arc GPUs, bringing the frame-boosting technique popularized by NVIDIA and AMD to Intel’s discrete and integrated graphics. The update means Panther Lake laptops with integrated Arc can now...
Meet the RTX $5,090 💸
The video spotlights MSI’s water‑cooled RTX 5090, now retailing at $5,090—a price tag that translates to roughly one dollar per pound of hardware. The host points out that this valuation rivals low‑end silver prices, underscoring how AI‑driven demand for RAM and GPUs...
Gemini 3 Deep Think: Optimizing 2D Semiconductor Fabrication
The video showcases Gemini 3’s Deep Think platform, specifically the DeepTank AI system, as a breakthrough tool for optimizing the fabrication of two‑dimensional (2D) semiconductors. By feeding a comprehensive thermal‑profile model into the growth furnace, the lab pushed crystal size from a...
Memory-Centric Computing - PACT 2022 Tutorial (08.10.2022)
The tutorial frames memory-centric computing as a response to rapidly growing data demands that are outpacing traditional compute-centric architectures. The speaker highlights that modern workloads—large neural networks, databases, graph analytics, and mobile applications—are increasingly bottlenecked by memory bandwidth, capacity, and...
Galaxy S25 Ultra — Still Worth Buying in 2026?
The video revisits Samsung’s Galaxy S25 Ultra a year after launch, assessing whether the flagship still merits purchase in 2026 as the S26 Ultra looms. Vick highlights the Gorilla Armor display’s scratch‑free durability, lightweight design, and the anti‑reflective coating that outshines...
Zettlab Ultra AI NAS Software Review
The reviewer tested the Zettlab Ultra (D6) AI NAS in a pre‑retail state, praising its automatic indexing of uploaded data and support for multiple local models and external GPUs, but criticizing limited user control over what gets crawled. The unit...
The Craziest GPU on the Planet..
MSI unveiled what it claims is the fastest RTX 5090 on Earth, a 2.5‑slot, 1,000‑watt‑limit graphics card that pushes the limits of consumer GPU design. The card packs a full‑coverage copper cold plate, carbon‑fiber shrouding and a 40‑phase VRM—far beyond the typical...
Alliwava G15 Disassembly & Teardown
The video walks through a full teardown of the Aliwava G15 laptop, highlighting its straightforward disassembly process. By removing eleven Phillips screws—two of which are captive near the hinge—and prying the bottom cover, users can access the battery, speaker cables,...
Logistics, 3D Printing, Lasers, Manufacturing - Openviro 2026 Update
Jeff provides a detailed status report on the Open VRO environmental sensor platform, focusing on the final manufacturing steps required before the first units ship. The video walks viewers through the 3D‑printed housing for the temperature‑humidity‑pressure probes, the laser‑engraved aluminum...
Monitor Names Suck so, so Much
PCWorld editor Brad Charkis lambastes the convoluted naming conventions for PC monitors, using MSI’s nearly indistinguishable model names—MSI MPG271QRQD-OLEDX50 versus MSI MAG272QPQD-OLEDX50—as a prime example. He argues that three-letter series like MEG, MAG and MPG add to buyer confusion, especially...
If This Is for You, It’s GREAT - Der8enchtable
Linus Media Group’s technical production assistant David Pancrrez walks viewers through the newly released Durbench test‑bench, a collaborative effort by Thermal Grizzly, Derbower and Elmer Labs that aims to modernize hardware‑testing platforms for 2026. The board integrates SATA bays, multiple fan...
Physical Data Transmission - Part 2: Amplitude Modulation & Frequency Shift Keying
The video introduces basic physical‑layer modulation, contrasting simple line codes with two more capable schemes—amplitude modulation (AM) and frequency shift keying (FSK). It explains why line codes waste bandwidth, carry few bits per symbol, and struggle on shared or non‑dedicated...
Moto G and Moto G Play (2026) Review: Which One to Get?
The video reviews Motorola's budget mid‑range Moto G 2026 and Moto G Play 2026, comparing specs, performance, and value at US price points. Both devices share a plastic vegan‑leather back, 4 GB RAM, 720p LCD with adaptive 90 Hz refresh, and a Dimensity 6300...
Lew Meets @Looirobot …
The video showcases Louie, a mobile AI robot marketed as a “life‑form” rather than a traditional product. Built around visual recognition and ChatGPT‑driven language models, Louie can perceive objects, people, and room layouts to deliver context‑aware responses. Louie’s voice interface activates...
Should You Buy an Apple Watch or an Oura Ring? ⌚️💍
The video pits Apple’s flagship smartwatch against the Oura Ring, asking which device delivers the best value for health‑focused consumers. The creator, who wears both simultaneously, frames the decision around functionality, battery life, and cost, noting that an ideal scenario...
STM32 USB MIDI Controller Tutorial (TinyUSB) - Phil's Lab #169
The video walks viewers through constructing a custom USB‑MIDI controller using an STM32H7 (SM3287) MCU and the TinyUSB stack. Phil repurposes a QueenB radio‑control PCB—originally not intended for music—to host thumb‑sticks, switches, and RGB LEDs, turning them into MIDI control‑change...
Is This PCB Layer Stackup OK? Solid GND – Signal Layer – Split Planes ???
The video examines how broken or split reference planes in a PCB stack‑up affect electromagnetic fields around high‑speed signal traces. Using Symbur simulations, the presenter compares a symmetric strip‑line with solid ground planes to configurations where one plane contains a...
Makera Z1: Precision CNC for Everyone
Makera’s Z1 desktop CNC smashed crowdfunding records, raising $10.62 million from over 7,000 backers and setting a new benchmark for consumer-grade machining. Priced from $899, the Z1 delivers 0.02 mm precision—three to five times finer than typical desktop CNCs—and can...
Cheapest GPUs and RAM - February 2026
The February 2026 installment of the “GPU and RAM price watch” series expands its scope to include SSDs, tracking U.S. retail prices via PC Part Picker. The host warns viewers that, contrary to expectations, component costs rose across the board,...
Why SSDs Are Getting Hotter | Solidigm
The video examines why modern solid‑state drives are running hotter than their predecessors, linking the trend to rising storage densities and the shift from PCIe Gen 4 to Gen 5 interfaces. As NAND cells pack more bits and controller logic speeds up, each...
Predicting Every Possible Network Path | Forward Networks
The video explains Forward Networks’ approach to predicting every possible network path by modeling load‑balancer behavior at a granular, packet‑by‑packet level. By applying queueing‑theory concepts, the team deconstructs policy‑based load‑balancing configurations to expose the underlying paths and decision points that...
Ethernet Switches in Space | Xsight Labs
Xsight Labs has its programmable, power-efficient Ethernet switches selected for SpaceX’s Starlink Gen 3 satellites, with multiple switches per satellite slated to launch in large volumes this year. The company won a competitive evaluation focused on programmability, power efficiency and...
Galaxy S26: What to Expect
Samsung is set to unveil its Galaxy S26 lineup on Wednesday, February 25, with shipments expected in early March. The company will keep pricing steady at $800 for the base S26, $1,000 for the Plus, and $1,300 for the Ultra,...
The New “AIR” Gaming Smartphone
Red Magic unveiled the 11 Air, its latest entry in the gaming‑phone segment, positioning it as a high‑performance device priced around $500. The handset arrives in two colors—Phantom black and Prism white—and targets mobile gamers seeking console‑like controls without a premium...
The Metadata Superpower You Need | Hammerspace
Hammerspace’s core capability is aggregating metadata from diverse underlying storage systems—object and NAS—into metadata servers that sit outside the data path to create a single global namespace. The platform assimilates metadata (a process that takes days, not instant) so organizations...
Make Your Data Storage Infinite | Hammerspace
Hammerspace offers a data-management platform that decouples data from underlying infrastructure, creating a virtually infinite, location-agnostic storage layer across clouds and on-prem systems. The company aggregates metadata across the data estate to eliminate silos, speed pipelines and enable seamless use...
Testing the UGREEN iDX6011 Pro AI NAS - REAL Turnkey Local AI LLM?
UGREEN’s iDX6011 Pro AI NAS is a crowdfunded network appliance aimed at bringing local LLM capabilities to on-premise data, with one-click model installs and mostly local processing but optional cloud integration. The reviewer finds the AI features still under development...
Why Generic AI Fails at Networking | Cisco Enterprise Networking
Cisco tested its proprietary deep networking model against generalist LLMs (GPT-5, Lambda/Gemma, GPTO OSS) on a 590-question MCQ benchmark and reported roughly a 20% performance advantage. The company credits that lead to training on its own networking data, yielding more...
Most Network Guys Aren't Ready For This | Cisco Data Center Networking
Cisco engineers warn that many enterprise network teams are unprepared for the unique demands of modern data-center traffic, particularly sustained, ultra-high-bandwidth flows between GPUs. Designing for these workloads requires precise spatial engineering of traffic, careful class-of-service configuration and detailed topology-aware...
The Hidden Problem With Cloud AI | Cisco Data Center Networking
Speakers argue that relying solely on rented cloud GPU resources is discouraging AI research because usage-based billing forces researchers to limit experimentation. They advocate for on-premises GPU infrastructure—capitalized once and reused over long lifecycles—to enable sustained exploration, hand down hardware...
You Don’t Need to Buy the Galaxy S25 Ultra. Here’s Why 📲
The video argues that consumers need not rush to purchase Samsung’s newest flagship, the Galaxy S25 Ultra, because a barely older model— the Galaxy S24 Ultra—offers virtually the same hardware at a dramatically lower cost on the secondary market. It...
Why Games Can’t Get Better Until 2030
The video, part of Harbor Unboxed Q&A, explores why gaming hardware improvements will stall until around 2030, focusing on current RAM and GPU shortages and how creators adapt. Tim and the host discuss DLSS 4.5's performance hit on legacy RTX cards,...
New DDR5 Refresh Modes on AM5 Motherboards
AMD AM5 motherboards from ASUS and Gigabyte now expose a Bank Refresh Mode setting that lets users choose between legacy (all-bank) refresh, fine-granularity refresh (FGR), and a mixed mode that switches dynamically. FGR refreshes one bank per bank group instead...