mikeayles

6 ms·

1.
by mikeayles 9d ago
The space is definitely heating up. Flux.ai being the incumbent Silixon based out of the UK have been showing demos Quilter and DeepPCB for layout The recent OpenAI demo... I went the other way and pre-designed sub circuits with a set of ru
2.
by mikeayles 1mo ago
Good question. short version: it's a shared wall, not a per-request bump. one KV260, one model on the fabric at a time, so the rotation is global. but the ~25s is mostly software, not fabric-dark time. the PL reconfigure is ~1-3s via f
3.
by mikeayles 1mo ago
I 100% agree the model is basically useless. that was never the deliverable. the artifact here is the inference engine, not the model living in it. 3.16M params at character level is just what fits in ~3MB of on-chip SRAM. Plus I treated it
4.
by mikeayles 1mo ago
Appreciated. and yeah, agreed the interesting comparison isn't "is this scalable as-is" (it isn't, PL fabric, cost, power), it's that the CIM / analog-compute startups you mention are chasing exactly this endpo
5.
by mikeayles 1mo ago
Just a quick test to see if i'm back
6.
by mikeayles 1mo ago
You're correct, the flat line is aggregate only. the fabric is saturated from a few dozen active clients onward, so extra connections can't buy throughput, they just queue. per-user p50/p95 across that same sweep: 17ms/3
7.
by mikeayles 1mo ago
I just posted a Show HN of my most recent side project, a live demo of a tiny-llm implemented in FPGA fabric, hitting an aggregate peak of 60,000tok/s, but a 'usable' model at 21,000tok/s Writeup and demo here: https:&#
8.
by mikeayles 1mo ago
I started this about 10 weeks ago when the Taalas chatjimmy demo first did the rounds, Llama 8B baked into custom silicon, 17k tok/s for a single user. Their whole thesis is that inference is bound by reading the weights, so stop fetch
9.
Show HN: A tiny LLM running at 21,000 tok/s on a $250 FPGA (Live Demo)(mikeayles.com)
79 points by mikeayles 1mo ago|33 comments
10.
by mikeayles 1mo ago
I'm using the AMD (Xilinx) K26. It's a Zynq Ultrascale+, the successor to the old classic 7000's. I'm running it on the KV260 dev board, because I'm using it for another side project. The K26 has a quad core A53 cor
11.
by mikeayles 1mo ago
It will be, I'm just not happy with my blog post before making it live. The blog will have a live interactive chat and a link to the repo with the HDL. I don't think anything I did was particularly novel, as I really just wanted t
12.
by mikeayles 1mo ago
AMD could have saved their money and used their own hardware! I've got a language model doing 60k tok/s on AMD hardware already, a Xilinx Kria K26 SOM, with the weights baked into URAM/BRAM with zero DRAM in the token loop. S
13.
Are you sure OneDrive has got your back(up)?(mikeayles.com)
2 points by mikeayles 3mo ago|0 comments
14.
Are you sure OneDrive has got your back(up)?(mikeayles.com)
3 points by mikeayles 3mo ago|0 comments
15.
by mikeayles 3mo ago
So for people wondering if it can be used to accelerate LLM inference, sadly not. I've been trying to hit 100,000tokens/s with a 3.28m dumb model, and even this is an order of magnitude too large to benefit. It appears to be focus
16.
by mikeayles 6mo ago
you may find this useful: https://phaestus.app/blog/blog0031 Edit: Forgot I also got doom running in openscad: https://www.mikeayles.com/blog/openscad-doom/ and doom running in openscad in th
17.
Does RAG Help AI Coding Tools?(mikeayles.com)
2 points by mikeayles 6mo ago|1 comments
18.
by mikeayles 6mo ago
I benchmarked Claude Code and GitHub Copilot on the same model (Haiku 4.5) with and without RAG-powered semantic search across 60 queries on a real codebase. RAG didn't make search more accurate on Claude Code, but it cut token consump
19.
Show HN: I reproduced the CL1 DOOM demo in 132 parameters(mikeayles.com)
1 points by mikeayles 6mo ago|0 comments
20.
by mikeayles 6mo ago
Rewriting the backend Bitwise Cloud, my semantic search for embedded systems docs Claude Code plugin from Python to Go. The problem was the ML dependencies. The backend uses BGE-small-en-v1.5 for embeddings and FAISS for vector search. Both
21.
by mikeayles 7mo ago
Fantastic comment, thanks! A review of the state of BEV's was actually going to be one of my next articles (hopefully after the additives). Are you happy for me to drop you an email to review a draft when I'm ready?
22.
by mikeayles 7mo ago
I do now: https://www.mikeayles.com/rss.xml
23.
by mikeayles 7mo ago
Good spot, and gruez is right about the caption too (fixed both, thanks). The car's L/hr figure was wrong. At 45 mpg (imperial) and 70 mph cruise, a car burns ~7 L/hr, not 3. That makes the flow rate ratio ~4x, which is consi
24.
by mikeayles 7mo ago
Probably the first thing to consider is the trucks have their speed calibrated periodically to ensure the accuracy of their tachographs (in the UK at least) so a truck doing 90kmph may show as 100kmph+ in a passenger car, I know my Volvo is
25.
by mikeayles 7mo ago
Sorry, got mixed up there, will amend, the 60 is for +3.5t! Edit: Nope, despite the vehicles only being able to propel themselves to 90kmph, the speed limit is indeed 60mph (in England and Wales, Scotland is a more sensible 56mph) https:&#
26.
by mikeayles 7mo ago
I have done a deep dive here: https://www.mikeayles.com/blog/on-vehicle-hydrogen-generatio... Short answer, it takes more energy to generate than the energy it produces. You can do things like only producing electrical
27.
by mikeayles 7mo ago
Fair point that a full TCO comparison would be more complete, and it's something I'm planning to cover in a later post. But the capital cost argument actually reinforces the conclusion rather than undermining it. Hydrogen fuel cel
28.
by mikeayles 7mo ago
Good catches, both of them. The hydrogen pathway should include grid transmission before the electrolyser, you're right. In practice it doesn't change the overall ratio much (multiplying by 0.95 on both sides) but the diagram shou
29.
CO2 Is the Wrong Number: Greenhouse Gas Equivalents for Road Freight(mikeayles.com)
1 points by mikeayles 7mo ago|0 comments
30.
Every Diesel Truck Has a Chemical Plant Bolted Underneath(mikeayles.com)
4 points by mikeayles 7mo ago|0 comments

More ›