cfg-0036
backfilled
citable URL: https://halobench.com/records/cfg-0036/ — this address never moves; the anchor /records/#cfg-0036 keeps resolving
| model | Qwen3.6-35B-A3B-UD-Q4_K_XL.gguf · UD-Q4_K_XL |
| engine | ggml-org/llama.cpp 3be50ccc2 · vulkan · host aihydra (igpu) |
| flags | -ngl 999 -fa 1 -b 2048 -ub 512 -ctk f16 -ctv f16 -t 16 |
| template | not recorded at test time |
| tree | upstream — stock |
aged evidence — reconstructed from the archive. Ingested from llama-bench, which measures throughput and not footprint: the memory block, chat-template hash and model revision are absent by construction. Peak memory comes from a Tier 0 capability probe, not from this tool. Everything present here — build commit, backend, KV quant, batch sizes, flash attention — is machine-reported.
Capability basis
unmeasured — performance numbers on this config stand on a guard alone, not an established capability
Runs on this config (10)
| run | date | kind | suite | metrics |
|---|---|---|---|---|
| run-0144 | 2026-08-12 | performance | llama-bench@10350 | prefill_tps=1251.48551 |
| run-0145 | 2026-08-12 | performance | llama-bench@10350 | decode_tps=61.83462 |
| run-0146 | 2026-08-12 | performance | llama-bench@10350 | prefill_tps=321.598303 |
| run-0147 | 2026-08-12 | performance | llama-bench@10350 | decode_tps=34.463715 |
| run-0148 | 2026-08-12 | performance | llama-bench@10350 | prefill_tps=705.35212 |
| run-0149 | 2026-08-12 | performance | llama-bench@10350 | decode_tps=51.020735 |
| run-0150 | 2026-08-12 | performance | llama-bench@10350 | prefill_tps=1149.512533 |
| run-0151 | 2026-08-12 | performance | llama-bench@10350 | decode_tps=59.873755 |
| run-0152 | 2026-08-12 | performance | llama-bench@10350 | prefill_tps=507.325549 |
| run-0153 | 2026-08-12 | performance | llama-bench@10350 | decode_tps=43.934256 |