Viewing profile — cpldcpu

its doing video, audio, images and motion. I think that counts as multimodal.

>When Intel finally shipped it at the 45-nanometer node in 2007, Gordon Moore called it the biggest change in transistor technology since the late 1960s. The breakthrough was not t…

hm.. has been quite a while for me. The good thing about the Tang Nano is that it is supported by the Yosys open source toolchain. There are quite a few resources on the web when y…

I had Opus 4.5 design an LLM inference engine in verilog, including firmware and automated verification a while ago: https://github.com/cpldcpu/smollm.c It's of course far from opt…

CISC view: Its another adressing mode. RISC view: SUBLEQ is already four instructions (2x memory access, alu, branch)

Great article! Yeah, that pattern can be seen everywhere in semiconductors. E.g. the transistor invention vs. Lilienfeld, Heil, Matare etc. So the scope is more narrow than "Invent…

Wow! And it also implements a very interesting variant of SUBLEQ that is turing complete. >This VM implements an OISC - a One Instruction Set Computer. That instruction takes three…

The early discovery of light emission from silicon carbide long before the first LEDs is a very interesting finding, worth pointing out. But alas, as ever so often, the article tur…

What do you mean with "open-source"? Of course, the inference code for all the open weight models is publically available - see llama.cpp or hf transformers. There are, however, ve…

Wow, nice! It's an honor. I guess nowadays one could use some of the 32bit WLCSCP microcontrollers to easily beat this.

More than 25 years ago, there was a show off, of building the smallest web server: https://web.archive.org/web/20000815063022/http://www-ccs.cs... Someone with an ACE1101 microcont…

From the blog post: https://github.com/wuxx/TD4-4BIT-CPU

Nice!

But why is this always the first comment on custom CPU builds? Can't there also be other designs out there?

nah... their founders and most of their employees are in france. Very ill-suited comparison to IBM.

https://huggingface.co/docs/inference-providers/pricing It's well buried though. Does not seem to be a focus of theirs.

Great to see this! Worth mentioning that Huggingface already offers a similar service. And they are also European: https://huggingface.co/docs/inference-providers/index https://hug…

I had to give it a try. Claude, the ole cheater, recognized what the file was, downloaded the psid from the web, found a wasm sid player and built a website around it: https://clau…

Yes, marks of AI all over the place. Also the SVGs. >No solution written, 100% score. Its weird. Turns out that hardest problem for LLMs to really tackle is long-form text.

Love it! Great idea for the dataset.

+1

They mentioned that they using strong quantization (iirc 3bit) and that the model was degradeted from that. Also, they don't have to use transistors to store the bits.

I wonder how well this works with MoE architectures? For dense LLMs, like llama-3.1-8B, you profit a lot from having all the weights available close to the actual multiply-accumula…