Bench m4 macstudio max 128g

[zhaopengme]

./ds4-bench 
-m ds4flash.gguf 
--prompt-file bench/promessi_sposi.txt 
--ctx-start 2048 
--ctx-max 65536 
--step-incr 2048 
--gen-tokens 128

m4 macstudio max 128g

[Copilot]

Pull request overview

Adds benchmark artifacts for an M4 Max Mac Studio (128GB) run of ds4-bench, extending the existing speed-bench/ hardware throughput dataset.

Changes:

• Added a new benchmark CSV with prefill/generation throughput across context sizes up to 65,536.
• Added a corresponding _ts.svg throughput plot for the new CSV.

Reviewed changes

Copilot reviewed 1 out of 2 changed files in this pull request and generated no comments.

File	Description
speed-bench/m4_macstudio_max_128g.csv	New throughput dataset (ctx vs prefill/gen t/s, kv-cache bytes).
speed-bench/m4_macstudio_max_128g_ts.svg	New SVG plot visualizing the dataset.

Comments suppressed due to low confidence (1)

speed-bench/m4_macstudio_max_128g.csv:35

• The CSV ends with extra blank lines. plot_speed.py uses csv.DictReader and will try to parse a blank record as a row (values become None), which can raise when converting ctx_tokens/prefill_tps/gen_tps. Please remove the trailing empty rows so the file ends after the last numeric line (optionally with a single final newline).

65536,2048,210.47,128,22.34,926033292

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.