dank/serverAppConfigs
1
0 Fork 0
Code Issues 4 Pull Requests Actions Packages Projects Releases Wiki Activity
906 Commits 67 Branches 8 Tags
c7db6d94e93e471e2ddb17345b4a832f8569f3e3
Commit Graph
4 Commits
Author SHA1 Message Date
Frank 9cf4b1de82 docs: qwen-fp8 endpoint published on 7773; fix log bullet placement 
PORTMAP=7773:7770 hook; health 200 on 127.0.0.1 and 192.168.0.235,
decode through the published path in band. 7773 chosen over 7770
(frank-primary owns 7770 when it returns). Log bullet recovered from a
misplaced insert into the 2026-09-28 cleanup section.
 2026-09-29 23:51:53 -07:00
Frank d28ab8ef8c docs: record the qwen-fp8-tuned-20260929 tag + digest-pinned image 
Repo tag qwen-fp8-tuned-20260929 at f532e03; serve script launches the
tcclaviger image by digest sha256:da402966... (local alias
qwen-fp8-runtime:tuned-20260929) — the floating :latest is no longer
referenced.
 2026-09-29 23:31:13 -07:00
Frank 57072a44df docs: final 256k verdict — boots at 4.03-4.5 GiB KV, band 34.1-36.5 straddles target 
The KV trim is inert: arena identical (5,648 slots) at 4.5 and 4.03 GiB
(planner buckets; the trim went to margin). Final 256k config
MAXLEN=262144 KV_BYTES=4328521728 prefetch off; full-speed alternative
MAXLEN=184224 at 3 GiB.
 2026-09-29 21:55:58 -07:00
Frank 0ccb01708b docs: qwen38-fp8 expert-offload tuning — prefetch off hits 37.4-40.6 tok/s (target met) 
Measured campaign on the ad hoc qwen-fp8 container (tcclaviger RDNA4
fork, 2x R9700, TP=2): telemetry (with_stack=false) measured the
prefetcher at 36% precision — ~half of gather traffic wasted; deleting
--expert-prefetch: 26.8 -> 40.57/37.44/37.57 tok/s. Prefetch-on shape
survived every other challenger (gather geometry, spec depth, TunableOp
[corrupts numerics], MNBT, concurrency, ngram [PLE fork bug]). 256k ctx
boots at 4.5 GiB KV (36.65/34.21/34.77, arena -14%). New runbook +
index links + log entry.
 2026-09-29 21:22:21 -07:00