Home › Evidence › Records › cfg-0048

cfg-0048

backfilled
citable URL: https://halobench.com/records/cfg-0048/ — this address never moves; the anchor /records/#cfg-0048 keeps resolving
modelQwen3.6-35B-A3B-UD-Q4_K_XL.gguf · UD-Q4_K_XL
engineggml-org/llama.cpp 3653e6d6d · vulkan · host aihydra (igpu)
flags-ngl 999 -fa 1 -b 2048 -ub 512 -ctk f16 -ctv f16 -t 16
templatenot recorded at test time
treeupstream — 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 (6)

rundatekindsuitemetrics
run-01962026-08-13performancellama-bench@1prefill_tps=277.396421
run-01972026-08-13performancellama-bench@1decode_tps=34.188152
run-01982026-08-13performancellama-bench@1prefill_tps=698.048959
run-01992026-08-13performancellama-bench@1decode_tps=50.313108
run-02002026-08-13performancellama-bench@1prefill_tps=489.314523
run-02012026-08-13performancellama-bench@1decode_tps=43.450871