# DeepSeek 1017: short-run EXL3 deployment and pool budget

The four selected C1 decode observations come from the raw quick-matrix CSVs, with 520-token thinking-off or 546-token effort-75 prompts and 64 actual output tokens. Each is one pass without variance estimates. Their first-to-last streamed-delta rate excludes initial prefill; end-to-end wall throughput remains a different field. This is not a long-context or sustained-response experiment.

- Type: Performance report
- Evidence: source_reported; site reproduced: false
- Model family: DeepSeek
- Model: LibertAIDAI/DeepSeek-V4.1-Flash-1017-EXL3-2.9bpw
- Original author: x00byte
- Original project: https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark
- Pinned document: https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark/blob/e85ed7ec3a403d8820a83e5aa33abe0985b58896/README.md
- Source revision: e85ed7ec3a403d8820a83e5aa33abe0985b58896
- Source commit time (not experiment date): 2026-10-17T18:27:21Z
- Checked: 2026-10-18
- Experiment date: None
- Runtime topics: TensorFold, vLLM
- Hardware topics: spark, nvidia
- Reported setup: {"system": "two DGX Spark GB10, 128 GB unified memory each, TP2, both available RoCE links", "node_count": 2, "parallelism": "TP2", "kv_dtype": null, "context_limit_tokens": 262144, "experiment_date": null}
- Performance figures: attributed selections with per-point conditions; no cross-source ranking

## Conditions

Unknown is null. Recipe options are not measured benchmark settings.

```json
{
  "conditions": {
    "model_revision": "afc7aa0aab98ee735c0f54e0a122019af1ac08f1",
    "weight_hash": null,
    "runtime_execution_revision": null,
    "input_tokens": null,
    "output_tokens": null,
    "concurrency": null,
    "cache_state": null,
    "quantization": null,
    "speculative_decoding": null,
    "temperature": null,
    "top_p": null
  },
  "test_profiles": [
    {
      "label": "tensorfold C1 thinking off",
      "evidence_url": "https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark/blob/e85ed7ec3a403d8820a83e5aa33abe0985b58896/results/tensorfold-quick.csv#L2",
      "conditions": {
        "model": "DeepSeek-V4.1-Flash-1017",
        "model_artifact": "LibertAIDAI/DeepSeek-V4.1-Flash-1017-EXL3-2.9bpw",
        "model_revision": "afc7aa0aab98ee735c0f54e0a122019af1ac08f1",
        "Engram_artifact": "deepseek-ai/DeepSeek-V4.1-Flash-1017",
        "Engram_revision": "1bae0e1088993b9b82de08f59de5be5a34e42170",
        "Engram_shard_SHA256": {
          "model-00047-of-00048.safetensors": "76a05106ece259be83d1bfea0dcb54f23377864ee60fd321690f35e6698a2af5",
          "model-00048-of-00048.safetensors": "9dc51433e195bc78b5ad846b68643cc0351d3b174e5e4309222d3b7402c83afd"
        },
        "weight_precision": "EXL3 2.9bpw; 6-bit dense MTP handling in vLLM overlay",
        "system": "two DGX Spark GB10, 128 GB unified memory each, TP2, both available RoCE links",
        "node_count": 2,
        "parallelism": "TP2",
        "experiment_date": null,
        "context_limit_tokens": 262144,
        "configured_max_requests": 8,
        "draft_depth_cap": 3,
        "concurrency": 1,
        "input_lines": 27,
        "output_tokens": 64,
        "max_output_tokens": 64,
        "repetitions": 1,
        "cache_state": "startup and earlier generation checks; no per-case warmup; TensorFold session/prepared caches disabled",
        "temperature": null,
        "top_p": null,
        "kv_dtype": null,
        "measurement_denominator": "(completion_tokens - 1) / first-to-last streamed delta interval; excludes initial prefill",
        "quality_evaluation": false,
        "long_response_steady_state_measured": false,
        "native_image_routing_in_text_matrix": false,
        "engine": "TensorFold Python 0.6.0 patched",
        "runtime_execution_revision": "c4646171139ee8a3c38103eaa1699dad226ec12b",
        "overlay_source_revision": "2dcce0303451da2a59f80e6c7413da8b3b0f7a57",
        "executed_image_digest": null,
        "thinking": false,
        "reasoning_effort": null,
        "input_tokens": 520,
        "actual_output_tokens": 64,
        "batch_wall_seconds": 2.4011132870218717,
        "end_to_end_output_wall_tps": 26.65430254620761,
        "TTFT_seconds": 1.313945789996069
      }
    },
    {
      "label": "tensorfold C1 thinking 75",
      "evidence_url": "https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark/blob/e85ed7ec3a403d8820a83e5aa33abe0985b58896/results/tensorfold-quick.csv#L10",
      "conditions": {
        "model": "DeepSeek-V4.1-Flash-1017",
        "model_artifact": "LibertAIDAI/DeepSeek-V4.1-Flash-1017-EXL3-2.9bpw",
        "model_revision": "afc7aa0aab98ee735c0f54e0a122019af1ac08f1",
        "Engram_artifact": "deepseek-ai/DeepSeek-V4.1-Flash-1017",
        "Engram_revision": "1bae0e1088993b9b82de08f59de5be5a34e42170",
        "Engram_shard_SHA256": {
          "model-00047-of-00048.safetensors": "76a05106ece259be83d1bfea0dcb54f23377864ee60fd321690f35e6698a2af5",
          "model-00048-of-00048.safetensors": "9dc51433e195bc78b5ad846b68643cc0351d3b174e5e4309222d3b7402c83afd"
        },
        "weight_precision": "EXL3 2.9bpw; 6-bit dense MTP handling in vLLM overlay",
        "system": "two DGX Spark GB10, 128 GB unified memory each, TP2, both available RoCE links",
        "node_count": 2,
        "parallelism": "TP2",
        "experiment_date": null,
        "context_limit_tokens": 262144,
        "configured_max_requests": 8,
        "draft_depth_cap": 3,
        "concurrency": 1,
        "input_lines": 27,
        "output_tokens": 64,
        "max_output_tokens": 64,
        "repetitions": 1,
        "cache_state": "startup and earlier generation checks; no per-case warmup; TensorFold session/prepared caches disabled",
        "temperature": null,
        "top_p": null,
        "kv_dtype": null,
        "measurement_denominator": "(completion_tokens - 1) / first-to-last streamed delta interval; excludes initial prefill",
        "quality_evaluation": false,
        "long_response_steady_state_measured": false,
        "native_image_routing_in_text_matrix": false,
        "engine": "TensorFold Python 0.6.0 patched",
        "runtime_execution_revision": "c4646171139ee8a3c38103eaa1699dad226ec12b",
        "overlay_source_revision": "2dcce0303451da2a59f80e6c7413da8b3b0f7a57",
        "executed_image_digest": null,
        "thinking": true,
        "reasoning_effort": 75,
        "input_tokens": 546,
        "actual_output_tokens": 64,
        "batch_wall_seconds": 2.141774678020738,
        "end_to_end_output_wall_tps": 29.881761446140466,
        "TTFT_seconds": 1.326543439005036
      }
    },
    {
      "label": "vllm C1 thinking off",
      "evidence_url": "https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark/blob/e85ed7ec3a403d8820a83e5aa33abe0985b58896/results/vllm-quick.csv#L2",
      "conditions": {
        "model": "DeepSeek-V4.1-Flash-1017",
        "model_artifact": "LibertAIDAI/DeepSeek-V4.1-Flash-1017-EXL3-2.9bpw",
        "model_revision": "afc7aa0aab98ee735c0f54e0a122019af1ac08f1",
        "Engram_artifact": "deepseek-ai/DeepSeek-V4.1-Flash-1017",
        "Engram_revision": "1bae0e1088993b9b82de08f59de5be5a34e42170",
        "Engram_shard_SHA256": {
          "model-00047-of-00048.safetensors": "76a05106ece259be83d1bfea0dcb54f23377864ee60fd321690f35e6698a2af5",
          "model-00048-of-00048.safetensors": "9dc51433e195bc78b5ad846b68643cc0351d3b174e5e4309222d3b7402c83afd"
        },
        "weight_precision": "EXL3 2.9bpw; 6-bit dense MTP handling in vLLM overlay",
        "system": "two DGX Spark GB10, 128 GB unified memory each, TP2, both available RoCE links",
        "node_count": 2,
        "parallelism": "TP2",
        "experiment_date": null,
        "context_limit_tokens": 262144,
        "configured_max_requests": 8,
        "draft_depth_cap": 3,
        "concurrency": 1,
        "input_lines": 27,
        "output_tokens": 64,
        "max_output_tokens": 64,
        "repetitions": 1,
        "cache_state": "startup and earlier generation checks; no per-case warmup; TensorFold session/prepared caches disabled",
        "temperature": null,
        "top_p": null,
        "kv_dtype": null,
        "measurement_denominator": "(completion_tokens - 1) / first-to-last streamed delta interval; excludes initial prefill",
        "quality_evaluation": false,
        "long_response_steady_state_measured": false,
        "native_image_routing_in_text_matrix": false,
        "engine": "vLLM + EXL3 overlay",
        "runtime_execution_revision": null,
        "overlay_source_revision": "6f7d1590ad49a2b8995188e45d7b9db31e677452",
        "executed_image_digest": "sha256:2f0cf3adc0f989c1d446be274df864eb799630175f604c3b22b71b7205971dce",
        "thinking": false,
        "reasoning_effort": null,
        "input_tokens": 520,
        "actual_output_tokens": 64,
        "batch_wall_seconds": 2.5652991679962724,
        "end_to_end_output_wall_tps": 24.948357212460998,
        "TTFT_seconds": 0.8555283379973844
      }
    },
    {
      "label": "vllm C1 thinking 75",
      "evidence_url": "https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark/blob/e85ed7ec3a403d8820a83e5aa33abe0985b58896/results/vllm-quick.csv#L10",
      "conditions": {
        "model": "DeepSeek-V4.1-Flash-1017",
        "model_artifact": "LibertAIDAI/DeepSeek-V4.1-Flash-1017-EXL3-2.9bpw",
        "model_revision": "afc7aa0aab98ee735c0f54e0a122019af1ac08f1",
        "Engram_artifact": "deepseek-ai/DeepSeek-V4.1-Flash-1017",
        "Engram_revision": "1bae0e1088993b9b82de08f59de5be5a34e42170",
        "Engram_shard_SHA256": {
          "model-00047-of-00048.safetensors": "76a05106ece259be83d1bfea0dcb54f23377864ee60fd321690f35e6698a2af5",
          "model-00048-of-00048.safetensors": "9dc51433e195bc78b5ad846b68643cc0351d3b174e5e4309222d3b7402c83afd"
        },
        "weight_precision": "EXL3 2.9bpw; 6-bit dense MTP handling in vLLM overlay",
        "system": "two DGX Spark GB10, 128 GB unified memory each, TP2, both available RoCE links",
        "node_count": 2,
        "parallelism": "TP2",
        "experiment_date": null,
        "context_limit_tokens": 262144,
        "configured_max_requests": 8,
        "draft_depth_cap": 3,
        "concurrency": 1,
        "input_lines": 27,
        "output_tokens": 64,
        "max_output_tokens": 64,
        "repetitions": 1,
        "cache_state": "startup and earlier generation checks; no per-case warmup; TensorFold session/prepared caches disabled",
        "temperature": null,
        "top_p": null,
        "kv_dtype": null,
        "measurement_denominator": "(completion_tokens - 1) / first-to-last streamed delta interval; excludes initial prefill",
        "quality_evaluation": false,
        "long_response_steady_state_measured": false,
        "native_image_routing_in_text_matrix": false,
        "engine": "vLLM + EXL3 overlay",
        "runtime_execution_revision": null,
        "overlay_source_revision": "6f7d1590ad49a2b8995188e45d7b9db31e677452",
        "executed_image_digest": "sha256:2f0cf3adc0f989c1d446be274df864eb799630175f604c3b22b71b7205971dce",
        "thinking": true,
        "reasoning_effort": 75,
        "input_tokens": 546,
        "actual_output_tokens": 64,
        "batch_wall_seconds": 2.4199117650277913,
        "end_to_end_output_wall_tps": 26.447245277666145,
        "TTFT_seconds": 0.8985815939959139
      }
    }
  ],
  "recipe_parameters": null,
  "deployment_options": []
}
```

## Reading notes

- A source document, not an independent experiment count. Preserve measurement type and unknown fields.

## Limits

- No independent Run AI Here model execution.
- Document revision is distinct from execution and weight identity.
- File bytes, host RSS, anonymous memory, MemAvailable, KV allocation and configured context are different quantities.

## License declaration

{"declared_spdx": null, "scope": "Source declaration only; base weights, converters, runtimes and source integration material retain their own terms."}


## Editorial answer

The four selected C1 decode observations come from the raw quick-matrix CSVs, with 520-token thinking-off or 546-token effort-75 prompts and 64 actual output tokens. Each is one pass without variance estimates. Their first-to-last streamed-delta rate excludes initial prefill; end-to-end wall throughput remains a different field. This is not a long-context or sustained-response experiment.

By Run AI Here editorial · 2026-10-18

Original source-grounded editorial synthesis with local Spark Qwen assistance and Codex review; no site model benchmark.

## What was measured?

The four selected C1 decode observations come from the raw quick-matrix CSVs, with 520-token thinking-off or 546-token effort-75 prompts and 64 actual output tokens. Each is one pass without variance estimates. Their first-to-last streamed-delta rate excludes initial prefill; end-to-end wall throughput remains a different field. This is not a long-context or sustained-response experiment.

Sources: https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark/blob/e85ed7ec3a403d8820a83e5aa33abe0985b58896/docs/benchmark-method.md, https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark/blob/e85ed7ec3a403d8820a83e5aa33abe0985b58896/results/tensorfold-quick.csv, https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark/blob/e85ed7ec3a403d8820a83e5aa33abe0985b58896/results/vllm-quick.csv

## Which files and memory controls belong to this profile?

Both runtimes use the same pinned EXL3 conversion and separately pinned official 1017 Engrams. The TensorFold Python profile has eight request slots, a shared 524288-token pool, a 262144 per-request ceiling and configured 5/4 GiB memory floors. Floors are admission controls, not observed peak memory or a proof of full-window concurrency.

Sources: https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark/blob/e85ed7ec3a403d8820a83e5aa33abe0985b58896/config/manifest.json, https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark/blob/e85ed7ec3a403d8820a83e5aa33abe0985b58896/docs/deployment.md, https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark/blob/e85ed7ec3a403d8820a83e5aa33abe0985b58896/config/tensorfold-knobs.env.example

## Can these rates be used for a VQ or million-token deployment?

No such transfer is established. The runtime, quantized tensors, request length and cache policy must stay attached to each observation. The text matrix predates the explicit native-image-routing flag; a later red-image smoke check confirms that path only. No comprehensive image or model-quality benchmark is claimed.

Sources: https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark/blob/e85ed7ec3a403d8820a83e5aa33abe0985b58896/docs/benchmark-method.md, https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark/blob/e85ed7ec3a403d8820a83e5aa33abe0985b58896/docs/deployment.md

## How does storage placement affect the deployment description?

The reported worker reads checkpoint files over a read-only NFS share, while active Engram lookups remain rank-local. The storage inventory includes separate source tables and packed files. Disk preparation, model residency and a shared KV budget must be recorded independently rather than adding all on-disk bytes into a claimed live memory requirement.

Sources: https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark/blob/e85ed7ec3a403d8820a83e5aa33abe0985b58896/docs/deployment.md

## Selected source-reported figures

Each point retains its own configuration, measurement unit and source. No site reproduction or cross-source ranking.

```json
[
  {
    "label": "tensorfold C1 thinking off",
    "metric": "decode_rate",
    "value": 58.05409336065285,
    "unit": "tokens/s",
    "statistic": "one pass; no run-to-run uncertainty estimate",
    "uncertainty": null,
    "conditions": {
      "model": "DeepSeek-V4.1-Flash-1017",
      "model_artifact": "LibertAIDAI/DeepSeek-V4.1-Flash-1017-EXL3-2.9bpw",
      "model_revision": "afc7aa0aab98ee735c0f54e0a122019af1ac08f1",
      "Engram_artifact": "deepseek-ai/DeepSeek-V4.1-Flash-1017",
      "Engram_revision": "1bae0e1088993b9b82de08f59de5be5a34e42170",
      "Engram_shard_SHA256": {
        "model-00047-of-00048.safetensors": "76a05106ece259be83d1bfea0dcb54f23377864ee60fd321690f35e6698a2af5",
        "model-00048-of-00048.safetensors": "9dc51433e195bc78b5ad846b68643cc0351d3b174e5e4309222d3b7402c83afd"
      },
      "weight_precision": "EXL3 2.9bpw; 6-bit dense MTP handling in vLLM overlay",
      "system": "two DGX Spark GB10, 128 GB unified memory each, TP2, both available RoCE links",
      "node_count": 2,
      "parallelism": "TP2",
      "experiment_date": null,
      "context_limit_tokens": 262144,
      "configured_max_requests": 8,
      "draft_depth_cap": 3,
      "concurrency": 1,
      "input_lines": 27,
      "output_tokens": 64,
      "max_output_tokens": 64,
      "repetitions": 1,
      "cache_state": "startup and earlier generation checks; no per-case warmup; TensorFold session/prepared caches disabled",
      "temperature": null,
      "top_p": null,
      "kv_dtype": null,
      "measurement_denominator": "(completion_tokens - 1) / first-to-last streamed delta interval; excludes initial prefill",
      "quality_evaluation": false,
      "long_response_steady_state_measured": false,
      "native_image_routing_in_text_matrix": false,
      "engine": "TensorFold Python 0.6.0 patched",
      "runtime_execution_revision": "c4646171139ee8a3c38103eaa1699dad226ec12b",
      "overlay_source_revision": "2dcce0303451da2a59f80e6c7413da8b3b0f7a57",
      "executed_image_digest": null,
      "thinking": false,
      "reasoning_effort": null,
      "input_tokens": 520,
      "actual_output_tokens": 64,
      "batch_wall_seconds": 2.4011132870218717,
      "end_to_end_output_wall_tps": 26.65430254620761,
      "TTFT_seconds": 1.313945789996069
    },
    "evidence_state": "source_reported",
    "site_reproduced": false,
    "source_document_id": "dsdual-results-tensorfold-quick.csv",
    "source_document_sha256": "3005b3be00581a962f9a269211eef2d598a6358d83c41e565ee1549b28e03c33",
    "source_line": 2,
    "evidence_url": "https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark/blob/e85ed7ec3a403d8820a83e5aa33abe0985b58896/results/tensorfold-quick.csv#L2"
  },
  {
    "label": "tensorfold C1 thinking 75",
    "metric": "decode_rate",
    "value": 77.49652509362433,
    "unit": "tokens/s",
    "statistic": "one pass; no run-to-run uncertainty estimate",
    "uncertainty": null,
    "conditions": {
      "model": "DeepSeek-V4.1-Flash-1017",
      "model_artifact": "LibertAIDAI/DeepSeek-V4.1-Flash-1017-EXL3-2.9bpw",
      "model_revision": "afc7aa0aab98ee735c0f54e0a122019af1ac08f1",
      "Engram_artifact": "deepseek-ai/DeepSeek-V4.1-Flash-1017",
      "Engram_revision": "1bae0e1088993b9b82de08f59de5be5a34e42170",
      "Engram_shard_SHA256": {
        "model-00047-of-00048.safetensors": "76a05106ece259be83d1bfea0dcb54f23377864ee60fd321690f35e6698a2af5",
        "model-00048-of-00048.safetensors": "9dc51433e195bc78b5ad846b68643cc0351d3b174e5e4309222d3b7402c83afd"
      },
      "weight_precision": "EXL3 2.9bpw; 6-bit dense MTP handling in vLLM overlay",
      "system": "two DGX Spark GB10, 128 GB unified memory each, TP2, both available RoCE links",
      "node_count": 2,
      "parallelism": "TP2",
      "experiment_date": null,
      "context_limit_tokens": 262144,
      "configured_max_requests": 8,
      "draft_depth_cap": 3,
      "concurrency": 1,
      "input_lines": 27,
      "output_tokens": 64,
      "max_output_tokens": 64,
      "repetitions": 1,
      "cache_state": "startup and earlier generation checks; no per-case warmup; TensorFold session/prepared caches disabled",
      "temperature": null,
      "top_p": null,
      "kv_dtype": null,
      "measurement_denominator": "(completion_tokens - 1) / first-to-last streamed delta interval; excludes initial prefill",
      "quality_evaluation": false,
      "long_response_steady_state_measured": false,
      "native_image_routing_in_text_matrix": false,
      "engine": "TensorFold Python 0.6.0 patched",
      "runtime_execution_revision": "c4646171139ee8a3c38103eaa1699dad226ec12b",
      "overlay_source_revision": "2dcce0303451da2a59f80e6c7413da8b3b0f7a57",
      "executed_image_digest": null,
      "thinking": true,
      "reasoning_effort": 75,
      "input_tokens": 546,
      "actual_output_tokens": 64,
      "batch_wall_seconds": 2.141774678020738,
      "end_to_end_output_wall_tps": 29.881761446140466,
      "TTFT_seconds": 1.326543439005036
    },
    "evidence_state": "source_reported",
    "site_reproduced": false,
    "source_document_id": "dsdual-results-tensorfold-quick.csv",
    "source_document_sha256": "3005b3be00581a962f9a269211eef2d598a6358d83c41e565ee1549b28e03c33",
    "source_line": 10,
    "evidence_url": "https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark/blob/e85ed7ec3a403d8820a83e5aa33abe0985b58896/results/tensorfold-quick.csv#L10"
  },
  {
    "label": "vllm C1 thinking off",
    "metric": "decode_rate",
    "value": 36.8753802051545,
    "unit": "tokens/s",
    "statistic": "one pass; no run-to-run uncertainty estimate",
    "uncertainty": null,
    "conditions": {
      "model": "DeepSeek-V4.1-Flash-1017",
      "model_artifact": "LibertAIDAI/DeepSeek-V4.1-Flash-1017-EXL3-2.9bpw",
      "model_revision": "afc7aa0aab98ee735c0f54e0a122019af1ac08f1",
      "Engram_artifact": "deepseek-ai/DeepSeek-V4.1-Flash-1017",
      "Engram_revision": "1bae0e1088993b9b82de08f59de5be5a34e42170",
      "Engram_shard_SHA256": {
        "model-00047-of-00048.safetensors": "76a05106ece259be83d1bfea0dcb54f23377864ee60fd321690f35e6698a2af5",
        "model-00048-of-00048.safetensors": "9dc51433e195bc78b5ad846b68643cc0351d3b174e5e4309222d3b7402c83afd"
      },
      "weight_precision": "EXL3 2.9bpw; 6-bit dense MTP handling in vLLM overlay",
      "system": "two DGX Spark GB10, 128 GB unified memory each, TP2, both available RoCE links",
      "node_count": 2,
      "parallelism": "TP2",
      "experiment_date": null,
      "context_limit_tokens": 262144,
      "configured_max_requests": 8,
      "draft_depth_cap": 3,
      "concurrency": 1,
      "input_lines": 27,
      "output_tokens": 64,
      "max_output_tokens": 64,
      "repetitions": 1,
      "cache_state": "startup and earlier generation checks; no per-case warmup; TensorFold session/prepared caches disabled",
      "temperature": null,
      "top_p": null,
      "kv_dtype": null,
      "measurement_denominator": "(completion_tokens - 1) / first-to-last streamed delta interval; excludes initial prefill",
      "quality_evaluation": false,
      "long_response_steady_state_measured": false,
      "native_image_routing_in_text_matrix": false,
      "engine": "vLLM + EXL3 overlay",
      "runtime_execution_revision": null,
      "overlay_source_revision": "6f7d1590ad49a2b8995188e45d7b9db31e677452",
      "executed_image_digest": "sha256:2f0cf3adc0f989c1d446be274df864eb799630175f604c3b22b71b7205971dce",
      "thinking": false,
      "reasoning_effort": null,
      "input_tokens": 520,
      "actual_output_tokens": 64,
      "batch_wall_seconds": 2.5652991679962724,
      "end_to_end_output_wall_tps": 24.948357212460998,
      "TTFT_seconds": 0.8555283379973844
    },
    "evidence_state": "source_reported",
    "site_reproduced": false,
    "source_document_id": "dsdual-results-vllm-quick.csv",
    "source_document_sha256": "befbf211d4f143798e12053ec21291cc77e0893284ccdb8d5a0aa19ef4722c6e",
    "source_line": 2,
    "evidence_url": "https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark/blob/e85ed7ec3a403d8820a83e5aa33abe0985b58896/results/vllm-quick.csv#L2"
  },
  {
    "label": "vllm C1 thinking 75",
    "metric": "decode_rate",
    "value": 41.447290817386964,
    "unit": "tokens/s",
    "statistic": "one pass; no run-to-run uncertainty estimate",
    "uncertainty": null,
    "conditions": {
      "model": "DeepSeek-V4.1-Flash-1017",
      "model_artifact": "LibertAIDAI/DeepSeek-V4.1-Flash-1017-EXL3-2.9bpw",
      "model_revision": "afc7aa0aab98ee735c0f54e0a122019af1ac08f1",
      "Engram_artifact": "deepseek-ai/DeepSeek-V4.1-Flash-1017",
      "Engram_revision": "1bae0e1088993b9b82de08f59de5be5a34e42170",
      "Engram_shard_SHA256": {
        "model-00047-of-00048.safetensors": "76a05106ece259be83d1bfea0dcb54f23377864ee60fd321690f35e6698a2af5",
        "model-00048-of-00048.safetensors": "9dc51433e195bc78b5ad846b68643cc0351d3b174e5e4309222d3b7402c83afd"
      },
      "weight_precision": "EXL3 2.9bpw; 6-bit dense MTP handling in vLLM overlay",
      "system": "two DGX Spark GB10, 128 GB unified memory each, TP2, both available RoCE links",
      "node_count": 2,
      "parallelism": "TP2",
      "experiment_date": null,
      "context_limit_tokens": 262144,
      "configured_max_requests": 8,
      "draft_depth_cap": 3,
      "concurrency": 1,
      "input_lines": 27,
      "output_tokens": 64,
      "max_output_tokens": 64,
      "repetitions": 1,
      "cache_state": "startup and earlier generation checks; no per-case warmup; TensorFold session/prepared caches disabled",
      "temperature": null,
      "top_p": null,
      "kv_dtype": null,
      "measurement_denominator": "(completion_tokens - 1) / first-to-last streamed delta interval; excludes initial prefill",
      "quality_evaluation": false,
      "long_response_steady_state_measured": false,
      "native_image_routing_in_text_matrix": false,
      "engine": "vLLM + EXL3 overlay",
      "runtime_execution_revision": null,
      "overlay_source_revision": "6f7d1590ad49a2b8995188e45d7b9db31e677452",
      "executed_image_digest": "sha256:2f0cf3adc0f989c1d446be274df864eb799630175f604c3b22b71b7205971dce",
      "thinking": true,
      "reasoning_effort": 75,
      "input_tokens": 546,
      "actual_output_tokens": 64,
      "batch_wall_seconds": 2.4199117650277913,
      "end_to_end_output_wall_tps": 26.447245277666145,
      "TTFT_seconds": 0.8985815939959139
    },
    "evidence_state": "source_reported",
    "site_reproduced": false,
    "source_document_id": "dsdual-results-vllm-quick.csv",
    "source_document_sha256": "befbf211d4f143798e12053ec21291cc77e0893284ccdb8d5a0aa19ef4722c6e",
    "source_line": 10,
    "evidence_url": "https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark/blob/e85ed7ec3a403d8820a83e5aa33abe0985b58896/results/vllm-quick.csv#L10"
  }
]
```

## Supplemental source identities

```json
{
  "supplemental_sources": [
    {
      "document_id": "dsdual-config-manifest.json",
      "parent_record_id": "deepseek-1017-dual-spark-quick",
      "publisher": "x00byte",
      "repository": "x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark",
      "revision": "e85ed7ec3a403d8820a83e5aa33abe0985b58896",
      "document_path": "config/manifest.json",
      "pinned_url": "https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark/blob/e85ed7ec3a403d8820a83e5aa33abe0985b58896/config/manifest.json",
      "document_sha256": "6b3da5443112253502befe2a866983a31d188af234aef5e03865a724dbe9b65a",
      "checked_on": "2026-10-18"
    },
    {
      "document_id": "dsdual-config-tensorfold-knobs.env.example",
      "parent_record_id": "deepseek-1017-dual-spark-quick",
      "publisher": "x00byte",
      "repository": "x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark",
      "revision": "e85ed7ec3a403d8820a83e5aa33abe0985b58896",
      "document_path": "config/tensorfold-knobs.env.example",
      "pinned_url": "https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark/blob/e85ed7ec3a403d8820a83e5aa33abe0985b58896/config/tensorfold-knobs.env.example",
      "document_sha256": "e26dc4761458be238d048c5a03b524993e7c632c27f8457527e3ba0030bc58c0",
      "checked_on": "2026-10-18"
    },
    {
      "document_id": "dsdual-config-vllm-knobs.env.example",
      "parent_record_id": "deepseek-1017-dual-spark-quick",
      "publisher": "x00byte",
      "repository": "x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark",
      "revision": "e85ed7ec3a403d8820a83e5aa33abe0985b58896",
      "document_path": "config/vllm-knobs.env.example",
      "pinned_url": "https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark/blob/e85ed7ec3a403d8820a83e5aa33abe0985b58896/config/vllm-knobs.env.example",
      "document_sha256": "ee4eb1535b45d28c47fdd5b136e4b04268a7955d2ac9ca83174afeb35c0d62d8",
      "checked_on": "2026-10-18"
    },
    {
      "document_id": "dsdual-docs-deployment.md",
      "parent_record_id": "deepseek-1017-dual-spark-quick",
      "publisher": "x00byte",
      "repository": "x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark",
      "revision": "e85ed7ec3a403d8820a83e5aa33abe0985b58896",
      "document_path": "docs/deployment.md",
      "pinned_url": "https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark/blob/e85ed7ec3a403d8820a83e5aa33abe0985b58896/docs/deployment.md",
      "document_sha256": "72fb210773604c15560e73d6db8ada144b0e953354201e42b3b7aebd4c018799",
      "checked_on": "2026-10-18"
    },
    {
      "document_id": "dsdual-docs-benchmark-method.md",
      "parent_record_id": "deepseek-1017-dual-spark-quick",
      "publisher": "x00byte",
      "repository": "x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark",
      "revision": "e85ed7ec3a403d8820a83e5aa33abe0985b58896",
      "document_path": "docs/benchmark-method.md",
      "pinned_url": "https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark/blob/e85ed7ec3a403d8820a83e5aa33abe0985b58896/docs/benchmark-method.md",
      "document_sha256": "33d83fe761ea2eaafaadee937fa212ffcec19d69a2528614776ad7fb1139ba3b",
      "checked_on": "2026-10-18"
    },
    {
      "document_id": "dsdual-results-README.md",
      "parent_record_id": "deepseek-1017-dual-spark-quick",
      "publisher": "x00byte",
      "repository": "x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark",
      "revision": "e85ed7ec3a403d8820a83e5aa33abe0985b58896",
      "document_path": "results/README.md",
      "pinned_url": "https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark/blob/e85ed7ec3a403d8820a83e5aa33abe0985b58896/results/README.md",
      "document_sha256": "380393cc1e86d175f21af1c216b613a2aad741ed90c851b7e8f491bed5231acf",
      "checked_on": "2026-10-18"
    },
    {
      "document_id": "dsdual-results-tensorfold-quick.csv",
      "parent_record_id": "deepseek-1017-dual-spark-quick",
      "publisher": "x00byte",
      "repository": "x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark",
      "revision": "e85ed7ec3a403d8820a83e5aa33abe0985b58896",
      "document_path": "results/tensorfold-quick.csv",
      "pinned_url": "https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark/blob/e85ed7ec3a403d8820a83e5aa33abe0985b58896/results/tensorfold-quick.csv",
      "document_sha256": "3005b3be00581a962f9a269211eef2d598a6358d83c41e565ee1549b28e03c33",
      "checked_on": "2026-10-18"
    },
    {
      "document_id": "dsdual-results-vllm-quick.csv",
      "parent_record_id": "deepseek-1017-dual-spark-quick",
      "publisher": "x00byte",
      "repository": "x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark",
      "revision": "e85ed7ec3a403d8820a83e5aa33abe0985b58896",
      "document_path": "results/vllm-quick.csv",
      "pinned_url": "https://github.com/x00byte/DeepSeek-V4.1-Flash-1017-Dual-Spark/blob/e85ed7ec3a403d8820a83e5aa33abe0985b58896/results/vllm-quick.csv",
      "document_sha256": "befbf211d4f143798e12053ec21291cc77e0893284ccdb8d5a0aa19ef4722c6e",
      "checked_on": "2026-10-18"
    }
  ]
}
```