$1,299 Intel Arc Pro B70 Hits Half of RTX 5090 Speed on Qwen3.8-27B After Extensive Tuning
maikerukonare · reddit · 2026-11-09
A Redditor benchmarked the 32GB Intel Arc Pro B70 ($1,299) against the RTX 5090 ($6,500+) running Qwen3.8-27B Q6: 52 vs 98 tok/s generation and 960 vs 2,830 tok/s prefill — half the speed for a fifth of the price ($41/GB vs $200+/GB).
Out of the box the card was terrible, but the author documented a full tuning guide:
- Kernel 6.17+, Resizable BAR on (needed a BIOS flash to disable CSM), Intel compute runtime from kobuk-team PPA
- Upgrading Mesa to 26.x alone gave 3x faster ollama Vulkan generation
- On Vulkan, keep KV cache at f16 — quantized KV (q80/q40) causes 1.75x–3x+ generation slowdowns
- Better: build llama.cpp with SYCL, then quantized KV works fine and MTP can be enabled (--spec-type draft-mtp --spec-draft-n-max 4)
- Vulkan only exposes 29GB by default; use --fit off with -c for the full 32GB
Test rig: Threadripper 1950X, 128GB DDR4, PCIe 3.0. Verdict: with 5090 prices inflated, the B70 is a compelling budget card for local LLM inference plus creative workloads.
More from Infra
- Jensen's downfall was expected: with Western ASICs rising, no more mercy needed — teortaxesTex · 2026-11-09
- Reflection AI Ships First Model Beam; Employee #1 Recalls a Year Building From 1k H100s to GB300 Clusters — bigblueboo · 2026-11-09
- BofA: Global data center spending tops $1T, projected to reach $2.1T by 2028 — Beth_Kindig · 2026-11-09
- RTX 3090 beats 128GB ASUS Spark and M5 Max on identical MiniMax H3 video benchmarks — Ok_Bird4483 · 2026-11-09
- Satire: Jensen's 'AI safety master plan' — resign, crash Nvidia, delay AGI a decade — ShakeelHashim · 2026-11-09
- Chamath: SpaceX could reprice Anthropic compute deal at 2-3x on 90-day notice — rohanpaul_ai · 2026-11-09