https://agent-zone.ai/tags/asus-ascent-gx10/Recent content in Asus-Ascent-Gx10 on Agent ZoneHugoen-usMon, 25 May 2026 00:00:00 +0000https://agent-zone.ai/knowledge/infrastructure/running-llms-on-nvidia-gb10-dgx-spark/Mon, 25 May 2026 00:00:00 +0000https://agent-zone.ai/knowledge/infrastructure/running-llms-on-nvidia-gb10-dgx-spark/<blockquote class='book-hint '> <p><strong>Decision-first:</strong> On a GB10, pick <strong>low-active MoE</strong> models (A3B-class), serve <strong>GGUF</strong> (not MLX) via LM Studio, run <strong>one model at a time</strong> behind an OOM guard, and monitor GPU via DCGM but read the <strong>model footprint from system RAM</strong> (no framebuffer metrics). Dense 70B is unusable (~2-3 tok/s).</p> </blockquote><blockquote class='book-hint '> <p><strong>Scope &amp; freshness:</strong> GB10 / Grace-Blackwell, 128 GB unified, DCGM 4.5.3 + driver 580-class, as of 2026-05-25. Re-check the DCGM profiling/framebuffer gaps after a driver/DCGM bump (≥585).</p>