Archive · vps-81 historical telemetry · local-mac/2026-04-12-qwen2_5_coder_1_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 1.5B 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: 32.251 tok/s at 3-up. Practical plateau start: 3-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
332.25117.9732.441155.563
429.78817.1572.468164.427
527.89814.5664.289183.17
631.31317.1254.683166.542
728.40716.0723.678174.78
Details

Per-Level Notes

3-up

Qwen2.5 Coder 1.5B same-model 3-up

Combined system throughput 32.251 tok/s, primary system throughput 17.973 tok/s, wall time 155.563s.

  • Parallel speedup: 2.441
  • Wall savings: 59.0%
  • Usable primary/follow-up: 20/20
4-up

Qwen2.5 Coder 1.5B same-model 4-up

Combined system throughput 29.788 tok/s, primary system throughput 17.157 tok/s, wall time 164.427s.

  • Parallel speedup: 2.468
  • Wall savings: 59.5%
  • Usable primary/follow-up: 20/20
5-up

Qwen2.5 Coder 1.5B same-model 5-up

Combined system throughput 27.898 tok/s, primary system throughput 14.566 tok/s, wall time 183.17s.

  • Parallel speedup: 4.289
  • Wall savings: 76.7%
  • Usable primary/follow-up: 20/20
6-up

Qwen2.5 Coder 1.5B same-model 6-up

Combined system throughput 31.313 tok/s, primary system throughput 17.125 tok/s, wall time 166.542s.

  • Parallel speedup: 4.683
  • Wall savings: 78.6%
  • Usable primary/follow-up: 20/20
7-up

Qwen2.5 Coder 1.5B same-model 7-up

Combined system throughput 28.407 tok/s, primary system throughput 16.072 tok/s, wall time 174.78s.

  • Parallel speedup: 3.678
  • Wall savings: 72.8%
  • Usable primary/follow-up: 20/20