Archive · vps-81 historical telemetry · local-mac/2026-04-12-qwen2_5_coder_0_5b-same-model-concurrency-sweep.html. Originally rendered 2026-04-12. Re-hosted from MyServers on 2026-05-06. Methodology and harness conventions may differ from what we use today; see /methodology.html for current standards. ← back to all benchmarks
Concurrency Sweep

Qwen2.5 Coder 0.5B Same-Model Sweep

This run selected the historically fastest local-Mac question types, repeated them into one focused packet, and then swept concurrency upward until throughput clearly stopped improving.

Peak combined system throughput: 87.773 tok/s at 5-up. Practical plateau start: 5-up.

Selection

Fast Question Types Chosen From History

QuestionCategoryAvg primary TPSAvg follow-up TPSAvg combined TPS
py_csv_parseparsing31.8926.9958.88
py_cli_argscli25.8226.6552.47
py_pydantic_modelvalidation25.7825.8551.62
py_typing_dataclasstyping26.2923.5549.84
py_file_scanfile_io23.8225.6749.49
Sweep

Concurrency Levels

ConcurrencyCombined system TPSPrimary system TPSParallel speedupTotal wall seconds
157.82142.8950.99861.172
258.3442.9011.51560.884
379.44259.1352.30144.762
460.34145.7722.58160.539
587.77364.9223.96142.143
659.49743.6933.38358.911
772.55952.3215.11350.056
Details

Per-Level Notes

1-up

Qwen2.5 Coder 0.5B same-model 1-up

Combined system throughput 57.821 tok/s, primary system throughput 42.895 tok/s, wall time 61.172s.

  • Parallel speedup: 0.998
  • Wall savings: -0.2%
  • Usable primary/follow-up: 20/20
2-up

Qwen2.5 Coder 0.5B same-model 2-up

Combined system throughput 58.34 tok/s, primary system throughput 42.901 tok/s, wall time 60.884s.

  • Parallel speedup: 1.515
  • Wall savings: 34.0%
  • Usable primary/follow-up: 20/20
3-up

Qwen2.5 Coder 0.5B same-model 3-up

Combined system throughput 79.442 tok/s, primary system throughput 59.135 tok/s, wall time 44.762s.

  • Parallel speedup: 2.301
  • Wall savings: 56.5%
  • Usable primary/follow-up: 20/20
4-up

Qwen2.5 Coder 0.5B same-model 4-up

Combined system throughput 60.341 tok/s, primary system throughput 45.772 tok/s, wall time 60.539s.

  • Parallel speedup: 2.581
  • Wall savings: 61.3%
  • Usable primary/follow-up: 20/20
5-up

Qwen2.5 Coder 0.5B same-model 5-up

Combined system throughput 87.773 tok/s, primary system throughput 64.922 tok/s, wall time 42.143s.

  • Parallel speedup: 3.961
  • Wall savings: 74.8%
  • Usable primary/follow-up: 20/20
6-up

Qwen2.5 Coder 0.5B same-model 6-up

Combined system throughput 59.497 tok/s, primary system throughput 43.693 tok/s, wall time 58.911s.

  • Parallel speedup: 3.383
  • Wall savings: 70.4%
  • Usable primary/follow-up: 20/20
7-up

Qwen2.5 Coder 0.5B same-model 7-up

Combined system throughput 72.559 tok/s, primary system throughput 52.321 tok/s, wall time 50.056s.

  • Parallel speedup: 5.113
  • Wall savings: 80.4%
  • Usable primary/follow-up: 20/20