romank

Well, never thought I would build a computer out of breadboards but here I am, dropbox.com/s/uilwgd3o2t55… :) Thank you @beneater for the fantastic videos and the kit!

It works 🚀 First demo of the new #386fpgacore running on real FPGA hardware (Sipeed Tang Console 138K). VGA output, 3DBench, Norton Commander, and Turbo C all running. Still slow and buggy — but a lot of programs already work.

Terry Tao understood Perelman's proof of the Poincare conjectore so you don't have to arxiv.org/pdf/math/06109…

@kateesackhoff @Seahawks So say we all!

Just landing California…AKA Enemy territory 🤣 always representing the Seahawks!!

78% pass rate on the real-mode 80386 instruction test set 🚀 I’d say I’m ~1/3 through real mode 386, factoring in debugging time. The 2,000+ lines of 80386 microcode is very interesting: compared to the 8086, there are dozens of internal registers (vs <10). 1/

...are somewhat missing the point: if one of these big problems turned out to be solvable using only existing mathematical insights and technology, it would be an immense disappointment: a wellspring that we previously thought to be gushing is, in fact, dry. (5/12)

Indeed, we identify such problems as being "interestingly hard" precisely because we intuit that they represent a "gap" in our current understanding and methods. This is why crackpots who claim to solve (say) the Riemann hypothesis using "tricks" or elementary methods... (4/12)

How many *true* theorems have plausibly written but erroneous proofs? These are much, much harder to catch. My guess is that it is a not insubstantial portion of the literature. 15/n, n=15.

as a word of advice to math-curious AI watchers, if someone actually resolves a Millenium problem (or really any significant mathematical problem), they will almost certainly be able to describe the main *ideas* of the proof, even prior to complete formalization of the argument

@IsaacKing314 What took you so long lol

Guys, I have even bigger news

Just uploaded #z8086 v0.2 🚀 — an 8086 FPGA core running the original microcode. This release adds the soc_hdmi demo (HDMI + UART + C firmware) and fixes a CPU bug. v0.2 shows that z8086 is now able to do useful things. GitHub: github.com/nand2mario/z80…

False. The change only seems radical because the new tools offer a potentially significant productivity boost. (Perhaps 20% when the dust settles.) This is not new. We’ve seen similar boosts many times over the last eight decades. The shifts from binary to assembler to G1 compilers (eg Fortran) to G2 (eg C) to G3 (VMs like Java) to G4 (eg Python). Radical change was predicted for each. FUD over programming jobs was rampant. Yet two things remained constant: * The ever increasing demand for programmers. * The principles of software engineering.

@awesomekling @FUTO_Tech Looking ripped, awesome!! The true OSS icon imo :)

Visiting @FUTO_Tech in the middle of winter bulking season. Thankfully they have a gym on site so I can get my shoulder work in 😄

A longer introduction to z8086 - the 8086 soft core that runs the original microcode - and some interesting discoveries building it: ☕⚡ 1. two important formulas in patent are wrong, 2. microcode greatly saves resource, 3. the 8086 interrupt bug. nand2mario.github.io/posts/2025/z80…

Still pondering a Christmas gift for the electronics enthusiast in your life? 🎁 @MouserElec has a limited number of @RadionaOrg ULX3S dev boards: #FPGA+ Espressif ESP32 (on the other side). More in stock early next year.

Hard times create keyboard users. Keyboard users create good times. Good times create mouse users. Mouse users create hard times.

@splinedrive Legendary!

Celebrating KianV RISC-V Chip Design

@NTDEV_ The price appears to be lots of swapping/page faults/lost performance

Ever wondered how low you can go? Here's Windows 11 running on just 184MB of RAM! Video down below

@bml_khubbard Great ToC, got the printed one. Thanks for spreading the knowledge :)

I’ve completed my year-long journey to becoming a published author in electrical engineering. Mastering FPGA Chip Design: For Speed, Area, Power, and Reliability is now part of the Elektor Academy Pro series and available in both print and eBook formats: elektor.com/products/maste…

@seizuresalad Ferry-ing, hopin’ ur doin’ great :)

Hi. How u doin'.

@PaulGradenwitz @konstructivizm Did the professor publish that or nothing of use came out of their efforts?

I had discussions with a professor who did exactly such simulations some years ago. They used Newton and not GR. That is why I stated it as a question. You can not let the expansion of space emerge out of the simulation but need to code it in the model. So, calculations can't be pure Newtonian because that assumes infinite interaction speed. That is why I wondered how they model it. But the 1/r² rule for mass is good enough for galaxy scale interaction.

Scientists have created the largest simulation of the Universe using 36,000 AMD Instinct MI250X graphics cards. This project will allow us to better understand the nature of dark matter and the processes of stellar evolution. In the video, each dot represents a separate galaxy. However, all this is only 0.001% of the scale model that is planned to be completed. On such a scale, the Earth turns out to be much smaller than we are used to thinking.