File size: 2,117 Bytes
f6bc13b | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 | # Measured results β GLM-5.3-Flash (NVFP4) + DFlash2, 2Γ DGX Spark (GB10), SGLang TP=2
First known deployment of this pairing on any hardware outside inco.ai (drafter repo
showed 0 downloads at deploy time; enabling PR #36708 merged ~7 h before first boot).
## Serving config (first-light envelope)
`mem-fraction-static 0.88`, `max-total-tokens 131072` (configurator resolved 64,832),
`mamba-full-memory-ratio 2`, context 65,536, `max_running_requests` resolved to 1,
KV bf16, draft tokens 8 (drafter-native block), draft attention flashinfer,
DSA tilelang retuned for GB10 (see patches/). Clocks: stock (no cap), engine warmed 2Γ800 tok.
## Decode throughput β warmed, temp 0, stream:false, 800 max_tokens, n=5, medians
| prompt | DFlash2 ON | DFLASH off (same stack) | ratio |
|---|---|---|---|
| code-1 (threaded queue impl) | **27.6 tok/s** (27.6β27.8) | 14.7 | **1.88Γ** |
| prose-1 (measurement essay) | **20.7 tok/s** (20.7β20.7) | 14.7 | **1.41Γ** |
Instantaneous scheduler gauge peaked at 45.5 tok/s during code decode; reported here
only as a gauge peak β sustained medians above are the comparable numbers.
## Acceptance (scheduler decode-batch log, during active code decode)
accept len 3.65β5.62 (of 9 per step: 7 drafts + verify + bonus), accept rate 0.38β0.66.
Consistent with the Qwen3.8 GB10 DFlash2 deploy's ~5/8 code acceptance.
## Losslessness (G6)
5 fixed prompts, temp 0, DFlash2-on vs DFLASH-off, token-for-token:
**NOT claimed lossless.** Of 5 prompts, 3 were VOID (harness captured only `content`,
which the auto-detected reasoning parser left empty β harness flaw, not a result),
1 matched exactly, 1 diverged at a formatting token (`**Step 1: Square each integer**`
vs `## Step 1: Square each number`) β a near-tie flipped by on/off-path numerics.
Full 20-prompt matrix with reasoning captured is queued. Treat DFlash2-on outputs as
distribution-preserving, not bit-identical, on this stack until that lands.
## Reference comparison (other published GB10 GLM-5.3 numbers, different stack)
- tonyd2wild vLLM TP2 + MTP-4: 21.8 tok/s median (peak 22.7); TP4: 35.7
|