{ "schema_version": 1, "builder_plan": { "scenario": "A self-hosted assistant is too expensive because its model uses most of the GPU memory.", "change": "Try the paper's supported lower-precision weights on a copy of your serving setup.", "baseline": "The same model at its original precision, on the same GPU.", "steps": [ "Locate the paper's implementation and check its license, model support, data needs, and hardware requirements. Stop here if you cannot obtain the required artifacts.", "Save a small set of representative product tasks and expected outcomes. Keep evaluation tasks out of any training or tuning, and choose quality and cost limits before running.", "Run the baseline and the proposed change on identical inputs. Keep other settings fixed and repeat paired runs to expose variation.", "Save per-task outputs, failures, timings, and all usage costs. Review regressions before expanding the test." ], "measure": "Peak memory, task pass rate, latency, and total serving cost at the same context and load.", "decision": "Keep the change only if it meets your preselected quality and latency limits and improves the metric you care about after setup and operating costs. Otherwise retain the baseline. This is an exploration plan, not a validated recommendation." }, "builder_review": null, "arxiv_id": "2609.21137", "title": "A Multi-Engine Dataflow for MoE Decoding on Scratchpad-Based Tensor Accelerators", "summary": "The method replaces dense BF16 expert-weight streaming during MoE decoding with a vector-quantized plus shared-basis low-rank representation, co-designed with a multi-engine dataflow that overlaps DMA transfers with compute on scratchpad-based accelerators. The cost is a custom weight conversion and a hardware-specific decoding pipeline tied to AWS Trainium3's scratchpad and multi-engine architecture, plus implementation complexity well beyond a standard inference stack. It can fail to transfer: the speedups depend on Trainium3's DMA and engine structure, the quality results are measured only against perplexity, and neither the representation nor the dataflow is available on a single 24 GB GPU or through third-party APIs.", "rationale": "The gains are real but bound to AWS Trainium3 hardware and custom kernel/dataflow engineering that the reader's single-GPU or API-based infrastructure cannot execute, so it is worth tracking rather than adopting.", "technique": "VQ plus low-rank expert representation with multi-engine MoE decode dataflow", "family": "quantizacao", "recommendation": "observar", "published": "2026-09-17", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2609.21137", "independent_implementations": 0, "total_implementations": 0, "stars": 0, "citations": 0, "claims": [ { "claim_id": "claim-01", "statement": "The method replaces dense BF16 expert-weight streaming during MoE decoding with a vector-quantized plus shared-basis low-rank representation, co-designed with a multi-engine dataflow that overlaps DMA transfers with compute on scratchpad-based accelerators. The cost is a custom weight conversion and a hardware-specific decoding pipeline tied to AWS Trainium3's scratchpad and multi-engine architecture, plus implementation complexity well beyond a standard inference stack. It can fail to transfer: the speedups depend on Trainium3's DMA and engine structure, the quality results are measured only against perplexity, and neither the representation nor the dataflow is available on a single 24 GB GPU or through third-party APIs.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2609.21137", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "1.15-1.31x speedup over AWS dense MoE megakernels at batch size 1, up to 1.7x at batch size 16, while matching or improving BF16-teacher perplexity across five MoE families on AWS Trainium3.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2609.21137", "source_page": null, "source_excerpt": "" } ], "exposure_map": [ { "dimension": "quality", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "compute", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "latency", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "operations", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "compatibility", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "security", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "data_and_training", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "reproducibility", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" } ], "risks": [], "minimum_test": [], "open_questions": [], "independent_tests": [], "lab_reproductions": [], "report_available": false, "validation_tier": "unknown", "evidence_tier": "unknown", "infrastructure_basis": "unknown", "equations": [], "equations_status": "not_fetched", "core_kind": "", "equations_fetched_at": "", "equations_source_url": "", "equations_source_sha256": "", "exposures_reviewed_at": "", "exposures_source_sha256": "" }