Prefill-parity gating of prefill tokens-per-second in cross-runtime reports
Parent: Mac local LLMs: Benchmarking and comparisons · Published reference · snapshot 2026-10-05
↓ Facts as markdownall context files
Prefill-parity gating of prefill tokens-per-second is already covered: llm-benchpacks computes a per-case parity status from prompt tokens and cached prompt tokens and prints prefill_tps only for comparable cases, as recorded in llm-benchpacks-workload-level-agent-benchmarks.md and time-to-first-...
These notes link each claim to its source. A source may be a research report hosted on this site rather than the primary document. A published reference means the content is available; it does not certify independent review or accuracy.Read the editorial policy and follow the sources before relying on a claim.
Facts
- Prefill-parity gating of prefill tokens-per-second is already covered: llm-benchpacks computes a per-case parity status from prompt tokens and cached prompt tokens and prints prefill_tps only for comparable cases, as recorded in llm-benchpacks-workload-level-agent-benchmarks.md and time-to-first-token-measurement-artifacts.md; the gate's inputs per server are listed in prompt-token-counting-parity-across-runtimes.md. [source]
Children
- No children recorded.