Follows the change already made on 4n4rch02. Updates the compose --port and
healthcheck plus every default in the bench tooling and the README, so the
scripts keep working without an explicit --port.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Counted the thinking block exactly via /tokenize. Server default stops it at
4095 tokens; a client sending reasoning_budget=16384 still gets 4095, so it
cannot be raised per request. Changing it means editing the compose file.
It truncates mid-derivation on a moderately hard problem, but the model
recovered and answered correctly (empty solution set, verified by brute
force), so there is no evidence 4096 is actually damaging output.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Verifying a K-token draft costs ~K times the CPU expert work, because each
token routes to its own 8-of-256 experts and nothing is amortized. The loss
is largest on structured code output (-51% for ngram-simple), i.e. exactly
where speculation should have won.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Hybrid SSM+attention MoE (qwen35moe): only 10 of 40 layers use full
attention, so 128k of KV costs ~1.3 GiB. The binding constraint is the
20.6 GiB of weights against 22 GiB of VRAM, handled with --n-cpu-moe.
Two findings drive the config:
- --n-cpu-moe strips experts from the first N layers, which -sm layer
assigns to CUDA0, so CUDA1 inherits every heavy layer and OOMs at any
offload level. -ts 24,16 rebalances it.
- --threads 8 (physical cores) beats 16 by 44% on generation; the expert
matmuls are bandwidth-bound and SMT siblings only contend.
Ships ncmoe=10 (53 t/s @8k, 34 t/s @97k) over the faster ncmoe=8 to keep
~1.3 GiB spare on CUDA0, which is shared with the desktop.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
DATAandClaude Opus 5