{
  "id": "spark-pair-moe-served-rate",
  "title": "Two-Spark MoE: served-rate and warmup audit",
  "kind": "benchmark",
  "source_type": "community_operator_report",
  "summary": "Two-node serving study that distinguishes end-to-end served rate from decode rate, and audits cache, warmup and checkpoint identity.",
  "runtime": "vLLM / SGLang",
  "runtime_topics": [
    "vLLM",
    "SGLang"
  ],
  "hardware_topics": [
    "spark",
    "nvidia"
  ],
  "platform_topics": [
    "linux"
  ],
  "model_label": "Qwen3.8-Flash-Next / GLM-5.3-Flash / DeepSeek-V4-Flash-0731",
  "model_family": "Multiple",
  "model_id": null,
  "model_release_date": null,
  "model_artifact": null,
  "evidence_state": "source_reported",
  "site_reproduced": false,
  "record_type": "curated_reference",
  "performance_ranking_allowed": false,
  "metrics": null,
  "metrics_status": "site_metrics_unmeasured; attributed_source_figures_in_editorial.reported_results",
  "conditions": {
    "model_revision": null,
    "weight_hash": null,
    "runtime_execution_revision": null,
    "input_tokens": null,
    "output_tokens": 512,
    "concurrency": null,
    "cache_state": null,
    "quantization": null,
    "speculative_decoding": null,
    "temperature": null,
    "top_p": null
  },
  "reported_setup": {
    "system": "2 x DGX Spark; 121 GiB unified memory each; 200 GbE",
    "node_count": 2,
    "parallelism": null,
    "kv_dtype": null,
    "context_limit_tokens": null,
    "experiment_date": null
  },
  "test_profiles": [
    {
      "label": "Fixed-length served rate",
      "evidence_url": "https://github.com/WeZZard/dgx-spark-bench/blob/25e8dbf1b6113dd0b3f12d27a670fb284944d260/README.md",
      "conditions": {
        "model_revision": null,
        "weight_hash": null,
        "runtime_execution_revision": null,
        "input_tokens": null,
        "output_tokens": 512,
        "concurrency": null,
        "cache_state": "full-shape warmup; per-node JIT cache; see row prefix-cache controls",
        "quantization": null,
        "speculative_decoding": "model-dependent NEXTN / DFlash2 / no-spec arms",
        "temperature": null,
        "top_p": null,
        "measurement_denominator": "whole request wall clock including TTFT"
      }
    }
  ],
  "deployment_options": [],
  "recipe_parameters": null,
  "measurement_scope": "inference_speed",
  "notes": [
    "Each request is forced to emit 512 tokens; served rate includes time to first token.",
    "Every result retains model, weights, KV type, engine, users and KV pool."
  ],
  "limitations": [
    "Text-0731 and Vision-Exp are separate checkpoints.",
    "A draft model consumes KV capacity; compare the exact per-row configuration."
  ],
  "comparison_status": "not_comparable",
  "editorial_status": "source_derived_detailed_article",
  "public_benchmark_snapshot": false,
  "source": {
    "publisher": "WeZZard",
    "url": "https://github.com/WeZZard/dgx-spark-bench",
    "repository": "WeZZard/dgx-spark-bench",
    "platform": "github",
    "revision": "25e8dbf1b6113dd0b3f12d27a670fb284944d260",
    "document_path": "README.md",
    "pinned_url": "https://github.com/WeZZard/dgx-spark-bench/blob/25e8dbf1b6113dd0b3f12d27a670fb284944d260/README.md",
    "report_url": "https://github.com/WeZZard/dgx-spark-bench/blob/25e8dbf1b6113dd0b3f12d27a670fb284944d260/README.md",
    "document_sha256": "b815d32807147f9638b1abbeadc02af257c55bd1f7651a123e5a8090bebe7469",
    "last_commit_at": "2026-09-06T00:35:29Z",
    "checked_on": "2026-10-17",
    "metadata_status": "checked"
  },
  "license": {
    "declared_spdx": "MIT",
    "scope": "Repository metadata only; per-file, model, data and conversion terms remain separate."
  },
  "editorial": {
    "version": 1,
    "published_on": "2026-10-17",
    "modified_on": "2026-10-17",
    "byline": "Run AI Here editorial",
    "creation_method": "AI-assisted source reading and original editorial synthesis; no independent model execution.",
    "answer": "The report divides all generated tokens by the complete request wall clock, including first-token delay. Every current table request is forced to 512 tokens with ignore_eos. Earlier naturally stopped outputs and 64-token correctness probes have different denominators and remain in the historical report rather than being merged into this cohort.",
    "sections": [
      {
        "heading": "What does served rate include?",
        "paragraphs": [
          "The report divides all generated tokens by the complete request wall clock, including first-token delay. Every current table request is forced to 512 tokens with ignore_eos. Earlier naturally stopped outputs and 64-token correctness probes have different denominators and remain in the historical report rather than being merged into this cohort."
        ],
        "citations": [
          "https://github.com/WeZZard/dgx-spark-bench/blob/25e8dbf1b6113dd0b3f12d27a670fb284944d260/README.md#L25"
        ]
      },
      {
        "heading": "Why were earlier results replaced?",
        "paragraphs": [
          "The author remeasured the campaign on 5 September after discovering a slow-clock GPU and unequal prefix-cache policy. The corrected runs disable prefix caching for all three models, exercise complete workload shapes during warmup and retain per-node JIT caches. Earlier measurements remain visible as history; they should not be treated as the corrected cohort."
        ],
        "citations": [
          "https://github.com/WeZZard/dgx-spark-bench/blob/25e8dbf1b6113dd0b3f12d27a670fb284944d260/README.md#L45"
        ]
      },
      {
        "heading": "What can the long-context rows establish?",
        "paragraphs": [
          "At one user in the long agentic workload, the reported medians are 22.5, 21.0 and 17.7 served tokens/s, with different between-campaign spreads. The author labels this row not separated by the data. Different draft models and KV pools remain attached to each model, and the high-concurrency rows do not even use the same number of users across models."
        ],
        "citations": [
          "https://github.com/WeZZard/dgx-spark-bench/blob/25e8dbf1b6113dd0b3f12d27a670fb284944d260/README.md#L70"
        ]
      },
      {
        "heading": "Why are failures and checkpoint identity useful?",
        "paragraphs": [
          "The README reports a Qwen FP8-KV configuration that failed on chunked input and GLM backend attempts that did not complete a forward pass. Its DeepSeek current rows use the text 0731 checkpoint, not Vision-Exp and not V4.1 Flash 1017. These failures and identities explain why a recipe label cannot substitute for the actual six-field serving configuration."
        ],
        "citations": [
          "https://github.com/WeZZard/dgx-spark-bench/blob/25e8dbf1b6113dd0b3f12d27a670fb284944d260/README.md#L50"
        ]
      }
    ],
    "facts": [],
    "reported_results": [
      {
        "label": "DeepSeek V4 Flash 0731, long agentic",
        "metric": "served_rate",
        "value": 22.5,
        "unit": "tokens/s",
        "statistic": "median of three campaigns",
        "uncertainty": {
          "spread_ratio": 1.04
        },
        "conditions": {
          "system": "2 x DGX Spark; 121 GiB unified memory each; 200 GbE",
          "node_count": 2,
          "parallelism": null,
          "kv_dtype": null,
          "context_limit_tokens": null,
          "experiment_date": "2026-09-05",
          "output_tokens": 512,
          "ignore_eos": true,
          "prefix_cache": false,
          "warmup": "all workload/user shapes, full generation",
          "repetitions": 3,
          "concurrency": 1,
          "measurement_denominator": "whole request wall clock including TTFT",
          "workload": "agentic 33k context",
          "model": "DeepSeek-V4-Flash-0731",
          "kv": "fp8_ds_mla",
          "kv_pool_tokens": 517484,
          "draft": "DSpark k5 probabilistic"
        },
        "evidence_state": "source_reported",
        "site_reproduced": false,
        "evidence_url": "https://github.com/WeZZard/dgx-spark-bench/blob/25e8dbf1b6113dd0b3f12d27a670fb284944d260/README.md#L70",
        "source_document_sha256": "b815d32807147f9638b1abbeadc02af257c55bd1f7651a123e5a8090bebe7469"
      },
      {
        "label": "Qwen3.8 Flash-Next, long agentic",
        "metric": "served_rate",
        "value": 21.0,
        "unit": "tokens/s",
        "statistic": "median of three campaigns",
        "uncertainty": {
          "spread_ratio": 1.01
        },
        "conditions": {
          "system": "2 x DGX Spark; 121 GiB unified memory each; 200 GbE",
          "node_count": 2,
          "parallelism": null,
          "kv_dtype": null,
          "context_limit_tokens": null,
          "experiment_date": "2026-09-05",
          "output_tokens": 512,
          "ignore_eos": true,
          "prefix_cache": false,
          "warmup": "all workload/user shapes, full generation",
          "repetitions": 3,
          "concurrency": 1,
          "measurement_denominator": "whole request wall clock including TTFT",
          "workload": "agentic 33k context",
          "model": "Qwen3.8-Flash-Next NVFP4",
          "kv": "BF16",
          "kv_pool_tokens": 161728,
          "draft": "NEXTN"
        },
        "evidence_state": "source_reported",
        "site_reproduced": false,
        "evidence_url": "https://github.com/WeZZard/dgx-spark-bench/blob/25e8dbf1b6113dd0b3f12d27a670fb284944d260/README.md#L70",
        "source_document_sha256": "b815d32807147f9638b1abbeadc02af257c55bd1f7651a123e5a8090bebe7469"
      },
      {
        "label": "GLM5.3 Flash, long agentic",
        "metric": "served_rate",
        "value": 17.7,
        "unit": "tokens/s",
        "statistic": "median of three campaigns",
        "uncertainty": {
          "spread_ratio": 1.32
        },
        "conditions": {
          "system": "2 x DGX Spark; 121 GiB unified memory each; 200 GbE",
          "node_count": 2,
          "parallelism": null,
          "kv_dtype": null,
          "context_limit_tokens": null,
          "experiment_date": "2026-09-05",
          "output_tokens": 512,
          "ignore_eos": true,
          "prefix_cache": false,
          "warmup": "all workload/user shapes, full generation",
          "repetitions": 3,
          "concurrency": 1,
          "measurement_denominator": "whole request wall clock including TTFT",
          "workload": "agentic 33k context",
          "model": "GLM-5.3-Flash",
          "kv": "BF16",
          "kv_pool_tokens": 162688,
          "draft": "DFlash2"
        },
        "evidence_state": "source_reported",
        "site_reproduced": false,
        "evidence_url": "https://github.com/WeZZard/dgx-spark-bench/blob/25e8dbf1b6113dd0b3f12d27a670fb284944d260/README.md#L70",
        "source_document_sha256": "b815d32807147f9638b1abbeadc02af257c55bd1f7651a123e5a8090bebe7469"
      }
    ],
    "source_document_sha256": "b815d32807147f9638b1abbeadc02af257c55bd1f7651a123e5a8090bebe7469"
  }
}