Tech News
← Home  ·  All topics

Graphics

14 GoKawiil briefs on this topic

Simple divide-by-depth formula explains 3D-to-2D graphics projection

A blog post highlighted by developer tsoding breaks down 3D perspective projection into one basic operation: dividing a point's x and y coordinates by its z (depth) value. The piece demonstrates this with live shader examples, showing how a ball's position and size shrink toward a vanishing point as it moves farther from the camera along the z-axis.

Developer builds FPGA clone of 3dfx Voodoo Graphics card paired with custom 486 CPU

An independent hardware developer has extended their z486_MiSTer project, a software-defined 486 PC built on FPGA hardware, by adding a working recreation of the original 3dfx Voodoo Graphics (SST-1) chip called zSST. Combined into a system dubbed z486 XL, running on a Xilinx KV260 board at 100 MHz, the setup can run period DOS games like Tomb Raider using their original 3dfx renderer.

Microsoft issues emergency fix for buggy Windows 11 25H2 update

Microsoft released an out-of-band patch to address problems introduced by its September 2026 Windows 11 update, which had shipped with new taskbar and Start menu customization features alongside nearly 1,000 security fixes. The rollout caused crashes on AMD Radeon GPUs, audio input/output failures, Remote Desktop instability, and Explorer black-screen issues after login. The emergency update addresses several of these regressions, though the AMD GPU crashes were not explicitly called out as fixed.

Rik Arends builds 3D flyable visualizer for 2.5M lines of code, plans to sell it

Makepad creator Rik Arends built a 3D navigable tool that renders up to 2.5 million lines of source code at over 120 frames per second, initially requiring 21GB of RAM before optimizations to indexing and streaming compression brought usage down to 3.5GB. He also tested loading Chromium's 51-million-line codebase, which took about 60 seconds.

MSI interrupt mode explained: why it steadies GPU performance, not speed

MSI (Message Signaled Interrupts) lets a device like a GPU notify the CPU of an interrupt by writing to a memory address rather than fighting for space on a shared physical hardware line. Windows 10 and 11 default to MSI-X, an expanded version introduced with PCI 3.0 in 2010, whenever a device supports it. This approach labels each interrupt with its source, avoiding the collisions that can occur on crowded shared IRQ lines.

Nvidia holds 90% share as desktop GPU shipments hit 12.5 million in Q2 2026

Jon Peddie Research reports desktop discrete graphics card shipments reached 12.5 million units in Q2 2026, the highest quarterly total since early 2022, even as prices climbed to record levels. Nvidia shipped about 11.25 million of those units for roughly 90% market share, while AMD shipped around one million cards for about 8%.

Explainer details how Logo builds programs from custom procedures

The piece walks through Logo's structure, showing how programmers define new procedures using 'to' and 'end' and combine simple shapes like squares into more complex ones like flowers and gardens. It notes that user-defined procedures behave identically to built-in primitives, so there's no visible distinction between them once written.

Nvidia's equity investments in AI firms surge to $99 billion

Nvidia's portfolio of equity stakes in tech and AI companies jumped more than tenfold over the past year, reaching $99 billion as of late July, up from roughly $7 billion a year earlier. The chip maker has committed over $40 billion to financing deals across the AI industry in 2026 alone, alongside separate pledges including up to $105 billion in conditional credit support for an OpenAI data center and a planned $12.9 billion acquisition of Hugging Face.

Bridson's 2007 one-page paper still powers modern Poisson disk sampling algorithms

The piece revisits Robert Bridson's 2007 single-page paper, which offers a fast algorithm for placing random points—like trees in a procedurally generated forest—so no two fall closer than a set minimum distance. It contrasts this compact, widely-cited solution with a nearly 1,000-page 2024 proof of the geometric Langlands conjecture, highlighting the range of scales in mathematical work. The algorithm uses a grid and an active-point list to sample points efficiently in an annulus around existing points, avoiding the slowdown of naive rejection sampling.

Standard RTX 5090 cards now retail for $5,090 amid relentless price surge

Nvidia's RTX 5090, originally launched at $1,999, is now selling for around $5,090 even in non-special-edition versions like the Asus TUF Gaming OC. Prices have jumped 20% in the past month and 41% over six months, putting the card 145% above its original MSRP. Even pricier variants exist, including the Asus ROG Astral RTX 5090 Edition 20 listed at $5,740.

GeForce GPUs Lead Fresh Wave of Global Price Hikes

New pricing data tracked across 10 countries shows graphics card prices jumping sharply in the six weeks since the last survey, reversing a modest decline seen earlier in 2026. Nvidia's GeForce lineup appears to be absorbing the steepest increases, based on lowest verified in-stock prices from comparison sites in each region.

Matrox traces its origins to 1976 Montreal startup by Trottier and Matić

Lorne Trottier and Branko Matić founded Matrox in Dorval, Quebec in 1976, combining their surnames to name the company. Their first product, the MTX-1632 video RAM board, generated a 32x16 ASCII display and sold $20,000 worth of orders after a free listing in Electronics magazine, making the fledgling company profitable almost immediately. They quickly followed up with the MTX-256, a raster graphics unit offering 256 by 256 dot resolution.