BTW, I find it pretty amazing that you can buy complete ESP32-S3-Zero boards on AliExpress for $5.33. That's for a little board with the chip, a ceramic antenna, a USB connector and an LED.
ESP32-C3-Zero are $2.80 each.
These prices are absolutely ridiculous for what you get. Also, the ease of use of these boards is incredible. Just connect them via USB to your computer and off you go.
That was my exact first computer with the CGA graphics, GEM Desktop, Locomotive Basic etc. All the software came in 4 360KB diskettes. The diskettes had each it's own beautiful color. There were 3 sample Basic programs that seemed to me quite advanced. One sample was a document management system. Another was a drawing program where you could draw with a mouse.
Didn't know that GEM Desktop was made by Digital Reasearch. It was a very good windowing system. I believe it was based on MS-DOS. There was also a CP/M OS.
So much software in only 1.2MB.
Pretty sure we have, e.g. [0], "TI's MSPM0Cx MCUs featuring Arm® Cortex®-M0+
core with 24 MHz CPU, 32-bit structure, 4 to 16
KB Flash provide cost-effective designs for electric
toothbrush and shaver applications." - compared to the 5150[1], it's 4x the bus width, 5x the CPU clock, albeit with fewer RAMs.
Probably, I was use this example because I feel youngers can't imagine how much we could do with such MS-DOS class computer, even within its limitations.
nit: 5.1MP/s to 11.2MP/s is not an increase of 220%, it's an increase of 120%. it's 2.2x faster, but percentage increase starts at 0%, not 100%. 100% faster is double, 200% faster is triple.
IME, they’re still pretty bad at getting specific implementation details of SIMD correct. The best I’ve done so far is to harness them to an actual piece of hardware and have them iterate until they see performance improvements but 1) Goodhart’s law, they tend to overfit the benchmark suite and 2) the resulting SIMD often becomes illegible.
For a lot of SIMD applications, overfitting isn't really a concern. In my use case I need to calculate the sum of absolute difference between two vectors for autocorrelation, there's no real possibility for overfitting, either it gets the correct answer or it doesn't. LLMs are good at reading through the CPU manual and trying different optimizations.
The resulting SIMD being illegible isn't an issue in my eyes, assembly isn't a language known for it's readability in general. If you want to really understand what the assembly is doing, ask your LLM to whip up a visualization and explanation tailored to your input data where you can see how each operation leads you to the final result.
Why is it still called SIMD-accelerated, if these are now just normal instructions. You just have to take a little bit more effort into restructuring your algorithms to make use of them.
ESP32-C3-Zero are $2.80 each.
These prices are absolutely ridiculous for what you get. Also, the ease of use of these boards is incredible. Just connect them via USB to your computer and off you go.
Indeed, don't forget it is more powerful than computers like this one,
https://en.wikipedia.org/wiki/Amstrad_PC1512
[0] https://www.ti.com/lit/ab/slaaeg0/slaaeg0.pdf
[1] https://en.wikipedia.org/wiki/IBM_Personal_Computer
The Pentium III was a superscalar core that scaled up to 1.4 GHz. This microcontroller is max 240 MHz and does less work per cycle.
So naturally an ESP32 can do so much more.
using faster with x/times is just begging for misunderstandings
[^1] Though, to be honest, x86 has great automatic SIMD tools so I never had to try too hard.
The resulting SIMD being illegible isn't an issue in my eyes, assembly isn't a language known for it's readability in general. If you want to really understand what the assembly is doing, ask your LLM to whip up a visualization and explanation tailored to your input data where you can see how each operation leads you to the final result.