Home › Evidence › Records › cfg-0034

cfg-0034

backfilled
citable URL: https://halobench.com/records/cfg-0034/ — this address never moves; the anchor /records/#cfg-0034 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 (10)

rundatekindsuitemetrics
run-01282026-08-12performancellama-bench@1prefill_tps=1110.521105
run-01292026-08-12performancellama-bench@1decode_tps=61.620189
run-01302026-08-12performancellama-bench@1prefill_tps=278.271174
run-01312026-08-12performancellama-bench@1decode_tps=34.24807
run-01322026-08-12performancellama-bench@1prefill_tps=697.733641
run-01332026-08-12performancellama-bench@1decode_tps=50.372403
run-01342026-08-12performancellama-bench@1prefill_tps=1020.010129
run-01352026-08-12performancellama-bench@1decode_tps=59.339747
run-01362026-08-12performancellama-bench@1prefill_tps=495.077024
run-01372026-08-12performancellama-bench@1decode_tps=43.464041