Candidate package · llama.cpp SYCL · 1× Intel Arc Pro B70

Qwen3.8 27B Q4_K_M on one Intel Arc Pro B70

Alibaba's Qwen3.8 27B on one Arc Pro B70 in 4-bit form with llama.cpp, no draft model. The lab's promoted one-card package, rebuilt from pinned source.

What we measured Lab-measured

27.8tok/s decode

Class-balanced median of per-input-class medians using conventional 99-interval rates on the fixed cold 12-prompt suite; target-only and cache-zero. The all-prompt median is 27.824790 tok/s. proof file

target-only = no draft model assisting · cache-zero = a fresh start, nothing pre-computed · 99-interval median = the middle rate across 99 measured stretches of a long answer

Model
Qwen3.8 27B
Publisher
Alibaba / Qwen
Checkpoint
ggml-org/Qwen3.8-27B-GGUF
Compression
Q4_K_M
Software
llama.cpp SYCL
Cards
1× Intel Arc Pro B70 32 GiB
Model weight bytes
19.0 GB
Operating systems
Linux
Delivery
native
Good for
general coding long context
Published
2026-08-22
Clean-host replay
not yet

Still missing before this becomes an install guide

  • tested clean-host Intel driver and oneAPI installation
  • clean-host source build and endpoint replay
  • batch-shape-invariant greedy output for multi-user serving

What to know

  • The beginner launch remains a conservative one-slot 8K profile. Separate audits measured exact one-slot HTTP active context through 32K and a stable output-audited 1→64 HTTP aggregate curve. The 64-slot profile nearly fills the card and multi-user greedy token identity is batch-shape-dependent, so it is a capacity result rather than a deterministic serving recommendation.
  • Qwen3.8 27B is dense; do not compare this one-card aggregate rate as if it were a sparse Qwen-derived MoE or an NVIDIA NVFP4 deployment.

Measured performance profiles Lab-measured

Q4_K_M raw decode with F16 KV Lab-measured

24.824.524.223.823.122.421.802K4K8K16K24K32KExisting context depth before tg128 · y: tok/s
Direct llama-bench raw-engine tg128 using the exact Q4_K_M TP1 lane, one B70, flash attention on, F16 K/V, and 5 repetitions at every displayed depth. This is a context-shape profile, not the 27.82 tok/s realistic-suite headline. Every marker is measured; no point is interpolated or extrapolated. evidence

Q4_K_M raw pp2048 with F16 KV Lab-measured

82592089385178071966802K4K8K16K24K32KExisting context depth before pp2048 · y: tok/s
Direct llama-bench raw-engine pp2048 using the exact Q4_K_M TP1 lane, one B70, flash attention on, F16 K/V, and 5 repetitions at every displayed depth. Every marker is measured; no point is interpolated or extrapolated. evidence

Q4_K_M raw decode with Q8_0 KV Lab-measured

24.322.421.118.714.912.410.702K4K8K16K24K32KExisting context depth before tg128 · y: tok/s
Alternative Q8_0 K/V operating profile on the same exact Q4_K_M TP1 lane, one B70, flash attention on, raw-engine tg128, and 5 repetitions per depth. It is shown separately because KV precision materially changes decode. Every marker is measured; no point is interpolated or extrapolated. evidence

Q4_K_M raw pp2048 with Q8_0 KV Lab-measured

81891288784377271166302K4K8K16K24K32KExisting context depth before pp2048 · y: tok/s
Alternative Q8_0 K/V operating profile on the same exact Q4_K_M TP1 lane, one B70, flash attention on, raw-engine pp2048, and 5 repetitions per depth. Every marker is measured; no point is interpolated or extrapolated. evidence

Qualified HTTP decode over exact active context Lab-measured

27.327.226.926.025.224.52K4K8K16K24K32KExact active prompt tokens · y: tok/s
One-slot llama-server HTTP completions on the exact accepted Q4_K_M TP1 stack, one B70, F16 KV, 33,024 configured context, cache zero, no truncation or context shift, and 128 returned token IDs. The registered repeated-token fixture is evidence grade C: it fixes context shape but is not natural prose. Every marker is measured; no point is interpolated or extrapolated. evidence

Qualified HTTP TTFT over exact active context Lab-measured

2,7755,57711,34323,47336,44050,2672K4K8K16K24K32KExact active prompt tokens · y: ms
TTFT from the same one-slot exact-token HTTP receipts as the qualified decode curve. Cache was zero and prompt counts were exact at every marker. The repeated-token fixture is evidence grade C and is not a natural-prose latency claim. Every marker is measured; no point is interpolated or extrapolated. evidence

Many people at once Lab-measured

Raw aggregate decode over parallel sequences Lab-measured

24.439.654.159.158.771.095.41248163264Concurrent engine sequences · log2 spacing · y: tok/s
Direct llama-batched-bench raw-engine continuous batching with independent pp128 prompts, interleaved tg256 decode, c32768, flash attention on, F16 KV, one B70, and the exact accepted package stack. This is a mechanism ceiling: it does not emit auditable completions and excludes HTTP, JSON, queueing, and server-scheduler overhead. No point is scaled, interpolated, or extrapolated. evidence

Output-audited HTTP aggregate decode Lab-measured

24.636.649.356.155.065.883.81248163264Concurrent HTTP users · log2 spacing · y: tok/s
llama.cpp native HTTP /completion on one B70 with 64 slots, 32K total F16 KV context, prompt cache and slot similarity disabled, unique short prompts, and 128 returned raw token IDs per request. Each marker is the median of two preregistered fresh-server attempts whose relative range was at most 2.03%. Output isolation passed with no cross-base oracle collision, but multi-user greedy token identity is batch-shape-dependent. No point is interpolated or extrapolated. evidence

How much faster could this get? Projected — not measured

The ML Bottleneck physics engine projects what stock software, a tuned run, and the physical ceiling look like for this exact model, compression, card count, and software. The grade is optimization headroom against the tuned-run target, not model quality.