Local LLM user weighs 128GB RAM upgrade on Z690 for longer contexts with Qwen3 8B
thatscoolbutno123 · reddit · 2026-11-05
A Reddit user runs Qwen3 8B Flash Next (IQ3XXS) locally on an older Z690 rig with an RX 9700 (32GB VRAM). With 4x16GB RAM, model usage already hits 55GB/64GB; they get 100 tps aggregated decode across 3 parallel sessions at 350k context (QK8V4). They're considering 4x32GB (€500-600, no heatspreaders but decent case airflow) to cache more sessions, use larger K-quants, and extend context without a big speed hit. Open questions: whether 2100-2600MHz RAM (much cheaper than ≥3200MHz) would noticeably hurt inference speed, and whether 150GB total VRAM+RAM can run something better than Qwen3 8B.
More from Infra
- Meta president says midterms will make building new US data centers 'very challenging' — ShakeelHashim · 2026-11-05
- Outsiders are signing $50M B300 cluster deals with zero space or power lined up — andersonbcdefg · 2026-11-05
- Local voice agent stack on RTX 3060 gives veteran user his 'magic' AI moment — 1or4s · 2026-11-05
- NASA makes history with first-ever quantum link between aircraft and ground — Polymarket · 2026-11-05
- Offline survival AI packs 1.5M articles into 8.8GB, works with no internet — mhdfaran · 2026-11-05
- LZ4, the extremely fast C compression algorithm, still pulling stars — lz4 · 2026-11-05