cfg-0184
fork
citable URL: https://halobench.com/records/cfg-0184/ — this address never moves; the anchor /records/#cfg-0184 keeps resolving
| model | Qwen3.8-27B-UD-Q4_K_XL.gguf · UD-Q4_K_XL · rev unsloth Dynamic V3.0 (downloaded 2026-09-03; file 17,559,178,144 bytes) |
| engine | ggml-org/llama.cpp c530ea79c753573c98796ebfaccb1aed8f5897b5 · rocm · host aihydra (igpu) |
| flags | -ngl 999 -fa on -c 262144 -b 2048 -ub 2048 -ctk f16 -ctv f16 --parallel 1 --slots --slot-save-path /var/lib/slotpin/kv-cache --cache-ram 8192 --reasoning-effort low --spec-type draft-mtp --spec-draft-n-max 2 --jinja --metrics --host 127.0.0.1 --port 5802 (fronted by the slotpin proxy on :8090, slots=1 max_inflight=1) |
| sampling | temperature 0 · max_tokens 256 |
| template | not recorded at test time |
| tree | fork — carrying con-0018 |
Capability basis
unmeasured — performance numbers on this config stand on a guard alone, not an established capability
Runs on this config (4)
| run | date | kind | suite | metrics |
|---|---|---|---|---|
| run-0661 | 2026-09-05 | performance | decode-depth-served@drive_depth_served | decode_tps=19.23 · prefill_tps=238.9 |
| run-0662 | 2026-09-05 | performance | decode-depth-served@drive_depth_served | decode_tps=16.99 · prefill_tps=141.4 |
| run-0663 | 2026-09-05 | performance | decode-depth-served@drive_depth_served | decode_tps=14.71 · prefill_tps=74.4 |
| run-0664 | 2026-09-05 | performance | decode-depth-served@drive_depth_served | decode_tps=12.49 · prefill_tps=49.7 |
Cited by — computed at build time, never stored
model pages qwen38-27b