Ken Shirriff

When lamenting the "old Internet", a lot of people forget that the vast majority of the people creating content on it were gainfully employed with strong career security. Meaning that they didn't need to make money from their hobbyist online projects, so they didn't need to monetize it.

People struggling to pay the bills couldn't afford to be on the old Web. There were as many (or more) desperately poor people in the world 30 years ago, but they were too poor to get online at all. The needed hardware and service were expensive, and online stuff was unimportant to everyday life. So the early Web was dominated by university students and people with middle class careers at established institutions.

Today it's a lot less expensive and a lot more universal to communicate electronically. Many desperate scrabblers now use electronic communications in hopes of making money -- whether it's offshore scammers trying to defraud American retirees or "mom bloggers" enticing you to click on their unoriginal and unneeded "awesome diaper brands" article.

In absolute terms there are still a lot of people who don't need to make money from writing and therefore write things that are as interesting and commercially unmotivated as the best of the 1990s Web. A couple of examples that frequently show up on HN:

Ken Shirriff's blog https://www.righto.com/

Daniel Lemire's blog https://lemire.me/blog/

There are also some bloggers and substack writers who do get income from writing, but write…

Hmm, a minor quibble: "VLSI" would be normally more than 10k gates or 100k transistors on a single chip, wouldn't it? But the 8008 had only about 3500 transistors (I don't see an exact count in http://www.righto.com/2016/12/die-photos-and-analysis-of_24.html) and so probably about 1000–1500 gates, so I think it should be called "LSI" rather than "VLSI".

A funny thing about the 8008 is that Intel's manual for its instruction set is unnecessarily shitty — even if you didn't know the history with Datapoint, the Intel manual is obviously not by the people who designed the instruction set because it's in hexadecimal, a tradition sadly followed by the 8080 and 8086 manuals. The Datapoint manuals, by contrast, are all in octal, making the machine code enormously easier to understand. (The H8 I grew up with used an Intel chip, but the front panel monitor program used octal.)

> 55 nanometers per transistor. Way off from the 2 nanometer feature size.

Cool. Following along here, I consider that transistors require multiple "features", and these days the more complicated transistor structure has enabled the increase in speed while lowering the power requirements. Power is also now a set of characteristics, the power needed to change transistor state, and the leakage current that remains when the transistor is not switching.

Not just a simple NPN junction FET anymore.

Then I think about those great microchip analyses on Ken Sheriff's blog. How very different transistor layouts show up in circuit designs. I can only imagine that modern high performance SOC design is even more complex.

https://righto.com

55 nanometers per transistor sounds like a useful number to me.

> currently there are actually no modern x86 CPUs on the market.

This feels like one of the stranger "no true Scotsman" arguments I've ever run across.

Even the 8086 had microcode that translated the instructions as generated by the compiler/programmer into the instructions that would be processed: http://www.righto.com/2022/11/how-8086-processors-microcode-engine.html

I would love to know what the person who wrote that stackexchange answer would say in response to the question, "so which x86 processors used 'real' x86 cores?" Because from the very beginning of x86, there's been translation of the front-door opcodes into internal opcodes via microcode.

If people are interested in the DX7's technical details, I've done some work reverse-engineering it, and some of Yamaha's other FM synths: https://ajxs.me/blog/tag/DX7.html

Ken Shirriff has also done some amazing work analysing the DX7's sound chips: https://www.righto.com/2021/11/reverse-engineering-yamaha-dx7.html

A really awesome documentary on the DX7 by MadFame: https://www.youtube.com/watch?v=sXt_NXjc7oY

There are a lot of software emulations of the DX7 that are quite realistic, like the amazing Dexed. Here's one that deserves a lot of credit for how accurate it is: https://github.com/chiaccona/VDX7

What is the board with a kajillion holes in it behind the esp32 called?

https://kn100.me/posts/desksniffer/pbfront.webp

It looks like it could come in handy for many esp32 projects.

p.s. In case the author comes across this comment - Absolutely superb write-up, your writing style reveals your curious and positive attitude, it's totally inspiring! I wish I could easily subscribe to future publications, like the emails I get from Ken Shirriff of https://righto.com. Also, your C++ is clear and easy to read, good job. Cheers.

Edit: @Klathmon, thanks for the info, that was very fast :)