Full-fabric VHDL LLM inference engine. Runs Qwen3.5-class transformer inference
llm.vhdl Full-fabric VHDL LLM inference engine. Runs Qwen3.5-class transformer inference (9B on a single card, 27B targeted across two) entirely in FPGA fabric: INT4 streaming matvec, Gated DeltaNet, gated attention, and a transformer sequencer, with the weights resident in on-card HBM. Outputs are validated against llama.cpp running the BF16 GGUF.
- ▪llm.vhdl Full-fabric VHDL LLM inference engine.
- ▪Runs Qwen3.5-class transformer inference (9B on a single card, 27B targeted across two) entirely in FPGA fabric: INT4 streaming matvec, Gated DeltaNet, gated attention, and a transformer sequencer, with the weights resident in on-card HBM.
- ▪Outputs are validated against llama.cpp running the BF16 GGUF.
Hacker News (AI / LLM) files mainly under ai. We currently carry 7,603 of its stories.
Story provenance
Source · retrieval · rights · ranking — open for full record
inspect →
Story provenance
Attribution is not the same as permission. This drawer separates discovery metadata, excerpts, WeSearch-generated summaries, reuse status, and whether the publisher receives the visit. Nothing here claims a legal grant the publisher has not made.
Record
| Original publisher | GitHub |
| Canonical URL | https://github.com/Nero7991/llm.vhdl |
| Publication time | Sun, 04 Oct 2026 20:48:43 +0000 |
| Retrieval time | 2026-10-04T21:37:08.977Z |
| Last seen | 2026-10-04T21:37:08.977Z |
| Headline source | Publisher (no WeSearch rewrite) |
| Excerpt source | publisher body |
| Excerpt method | First ~120 words (~800 chars) of extracted publisher body, fair-use limited. |
| Summary | WeSearch · cerebras-chat (WeSearch summarizer) |
| Summary source text | contentText |
| Citation coverage | Summary is a WeSearch-generated derivative; primary citation is the original publisher URL. |
| Cluster | ucNQngCdOsGS · 1 stories |
| Cluster logic | Grouped by semantic title/content similarity across sources within a rolling window. Same-publisher template collisions are excluded from coverage comparison. |
| Ranking reason | Story pages are not engagement-ranked. Hub feeds use recency, with optional source-diversified chronological ordering (cap consecutive stories per source). No personalized ranking. |
| Publisher visit | Yes — open original |
| Substitutes article? | No — link-out required for full text |
Rights status (four layers)
WeSearch handling by dimension
| Indexing | May the item be indexed (stored, ranked, made findable)? | Allowed |
| Snippet | May a short excerpt of the publisher's text be shown? | Allowed |
| AI summary | May WeSearch generate its own short summary of the article? | Limited |
| Retrieval / RAG | May the content be exposed for third-party retrieval-augmented generation? | Not asserted |
| Model training | May the content be used to train AI models? | Not asserted |
| Commercial reuse | May the content be reused commercially? | Not permitted |
Basis: Derived from the published RSS/Atom feed. Contact: [email protected]. Reviewed: 2026-07-24.
Opening excerpt (first ~120 words) tap to expand
llm.vhdl Full-fabric VHDL LLM inference engine. Runs Qwen3.5-class transformer inference (9B on a single card, 27B targeted across two) entirely in FPGA fabric: INT4 streaming matvec, Gated DeltaNet, gated attention, and a transformer sequencer, with the weights resident in on-card HBM. Outputs are validated against llama.cpp running the BF16 GGUF. Its per-layer activations, captured through llama.cpp's eval callback, are compared with a bit-accurate C model of the INT4 / fixed-point datapath and with residuals read back from the card, and the tokenizer and chat template are bit-exact against llama.cpp. Details: tools/ref9b/README.md. Target cards: SQRL FK33 (xcvu33p-fsvh2104-2L-e, 8 GiB HBM, PCIe Gen3 x4) and Jungle Cat (2x xcvu35p modules, 8 GiB HBM each, Ethernet only, no PCIe).
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at GitHub.