llama.cpp dflash2: Qwen 3.8 27B Inference Speed Up to 3x
Top-Eye-8104 · reddit · 2026-08-20
A user tested the new dflash2 feature in llama.cpp (PR #27342) on an RTX 6000 using Qwen 3.8 27B, comparing four decoding strategies:
Median Results (4 tasks):
- Baseline: 47.4 tok/s
- MTP: 114.7 tok/s
- DFlash: 99.3 tok/s
- DFlash2: 140.6 tok/s
DFlash2 delivers an average 3x speedup, though performance varies by task, with one test showing only a 1.5x gain.
Related event: llama.cpp's New DFlash2 Boosts Qwen Inference Speed Up to 3x(2 posts)→
More from Infra
- Prediction: Opus 5.5-level models on laptops by 2027, backed by Qwen 27B data — cephaloform · 2026-10-16
- Oracle down 12% in a week, nearly 50% off June high as OpenAI exposure spooks market — alvelda · 2026-10-16
- US data center moratoriums top 400 — an AI bull explains why they're actually good for AI — kevinsxu · 2026-10-16
- Microsoft Research unveils Magnetar, the fastest zkVM at 10-39x speedups and post-quantum security — jedisct1 · 2026-10-16
- Bonsai 2 2B hits 120 tok/s on iPhone 18 Pro, outpaces Qwen 3.5 2B at half the memory — cephaloform · 2026-10-16
- Samsung cuts Q4 phone output as Nittobo shifts E-glass to AI-chip T-glass — zephyr_z9 · 2026-10-16