Frank cadc7f2768 frank-default-large: threads-batch=16 confirmed optimal + decode benchmark results ...

Decode sweep (TB=8,16,32, 8K+ prompt, 256 gen tokens):
  TB=8:  3.3 tok/s (76.6s)
  TB=16: 3.4 tok/s (74.8s)
  TB=32: 3.4 tok/s (75.0s)

Decode is flat; TB=16 wins on prefill (5.1% faster) with zero decode tradeoff.

Also:
- Fix :: syntax error in serve-frank-large.sh
- Document systemctl disable destroys unit file; use mask/unmask instead

2026-08-27 13:20:52 -07:00

..

.obsidian

Clarify llama-bench harness scope

2026-08-25 19:07:46 -07:00

Architecture

Document canonical Caddy migration

2026-08-25 21:59:59 -07:00

Dankclaw

caddy + pihole: make both boot-persistent and self-healing

2026-08-26 19:52:57 -07:00

Dankserver

Add CLI helper for entropy volume

2026-08-25 15:38:57 -07:00

Infrastructure

caddy: drop the systemctl enable; the quadlet generator already enables it

2026-08-26 20:19:03 -07:00

LLM

frank-default-large: threads-batch=16 confirmed optimal + decode benchmark results

2026-08-27 13:20:52 -07:00

Research

docs(kv): add bit-density KV cache quant table across 3 runtimes

2026-08-25 20:03:13 -07:00

index.md

docs: link STRIX runbook from top-level index

2026-08-27 03:01:27 -07:00

log.md

caddy: drop the systemctl enable; the quadlet generator already enables it

2026-08-26 20:19:03 -07:00