THE SOURCE IN CONTEXT
Boot 10 uses four Sparks at TP4 with node-local disk-backed Engram rows, DSpark k5, tools and vision enabled, 300K configured context and GPU memory utilization 0.80. Later EXL3, TP3 and longer-context boots in the same repository are separate experiments. This page keeps the boot-10 measurements anchored to their own section.
Which boot does this article cover?
Boot 10 uses four Sparks at TP4 with node-local disk-backed Engram rows, DSpark k5, tools and vision enabled, 300K configured context and GPU memory utilization 0.80. Later EXL3, TP3 and longer-context boots in the same repository are separate experiments. This page keeps the boot-10 measurements anchored to their own section.
Source: tonyd2wild · pinned document
Why are aggregate and decode rates different?
The author counts tokens from the server usage field rather than stream chunks. Aggregate output divides all stream tokens by batch wall time, including the wait for the first token. Per-stream decode counts tokens after the first over time after that first token. At C1 the category means are consequently 37.95 aggregate and 43.12 decode tokens/s, with 0.441-second mean TTFT.
Source: tonyd2wild · pinned document
How repeatable is the published sweep?
The fixed category prompt set is warmed, sampled at temperature zero with thinking off, and released in concurrent batches. Each cell has one batch. The report warns of a GB10 slow-clock state that can move an individual cell substantially. Its counting ceiling is excluded from the eight-category mean, and should stay excluded when quoting that mean.
Source: tonyd2wild · pinned document
What can the data be used for?
Use the report to understand this TP4 configuration and its content-dependent latency. It does not establish 1017 performance, a broad workload average or a controlled comparison with later engine lanes. The prefill table uses cold unique prompts and one-token replies, a different profile from the short streaming sweep.
Source: tonyd2wild · pinned document
Reported figures and their conditions
Selected facts reported by the original author. Each row has its own configuration and measurement boundary. Run AI Here has not reproduced these tests.
| Configuration / workload | Metric | Reported value | Evidence |
|---|---|---|---|
| Eight-category mean, C1 | aggregate output | 37.95 tokens/smean across eight categories; counting excluded | Original row ↗ |
| Eight-category mean, C1 | per stream decode | 43.12 tokens/smean across eight categories; counting excluded | Original row ↗ |
| Eight-category mean, C1 | ttft | 0.441 secondsmean across eight categories | Original row ↗ |
| Eight-category mean, C6 | aggregate output | 131.86 tokens/smean across eight categories; counting excluded | Original row ↗ |
Eight-category mean, C1 · aggregate output · conditions
| system | 4 x NVIDIA DGX Spark |
|---|---|
| node_count | 4 |
| parallelism | TP4 |
| kv_dtype | Not specified |
| context_limit_tokens | 300000 |
| experiment_date | Not specified |
| model | DeepSeek V4.1 Flash launch checkpoint, not 1017 |
| speculative_decoding | DSpark k5 |
| temperature | 0 |
| thinking | False |
| warmup | performed |
| repetitions | one batch per cell |
| input_tokens | about 30–120; summary prompt longer |
| output_token_budget | 150–256; natural stopping |
| weight_precision | MXFP4 experts / MXFP8 dense |
| concurrency | 1 |
Eight-category mean, C1 · per stream decode · conditions
| system | 4 x NVIDIA DGX Spark |
|---|---|
| node_count | 4 |
| parallelism | TP4 |
| kv_dtype | Not specified |
| context_limit_tokens | 300000 |
| experiment_date | Not specified |
| model | DeepSeek V4.1 Flash launch checkpoint, not 1017 |
| speculative_decoding | DSpark k5 |
| temperature | 0 |
| thinking | False |
| warmup | performed |
| repetitions | one batch per cell |
| input_tokens | about 30–120; summary prompt longer |
| output_token_budget | 150–256; natural stopping |
| weight_precision | MXFP4 experts / MXFP8 dense |
| concurrency | 1 |
| denominator | after first token |
Eight-category mean, C1 · ttft · conditions
| system | 4 x NVIDIA DGX Spark |
|---|---|
| node_count | 4 |
| parallelism | TP4 |
| kv_dtype | Not specified |
| context_limit_tokens | 300000 |
| experiment_date | Not specified |
| model | DeepSeek V4.1 Flash launch checkpoint, not 1017 |
| speculative_decoding | DSpark k5 |
| temperature | 0 |
| thinking | False |
| warmup | performed |
| repetitions | one batch per cell |
| input_tokens | about 30–120; summary prompt longer |
| output_token_budget | 150–256; natural stopping |
| weight_precision | MXFP4 experts / MXFP8 dense |
| concurrency | 1 |
Eight-category mean, C6 · aggregate output · conditions
| system | 4 x NVIDIA DGX Spark |
|---|---|
| node_count | 4 |
| parallelism | TP4 |
| kv_dtype | Not specified |
| context_limit_tokens | 300000 |
| experiment_date | Not specified |
| model | DeepSeek V4.1 Flash launch checkpoint, not 1017 |
| speculative_decoding | DSpark k5 |
| temperature | 0 |
| thinking | False |
| warmup | performed |
| repetitions | one batch per cell |
| input_tokens | about 30–120; summary prompt longer |
| output_token_budget | 150–256; natural stopping |
| weight_precision | MXFP4 experts / MXFP8 dense |
| concurrency | 6 |
AI-assisted source reading and original editorial synthesis; no independent model execution.