8GB VRAM runs Flux and Wan 2.2 locally fine: three wrong settings, not the GPU, were the bottleneck
leonbuilds · reddit · 2026-09-08
On an 8GB RTX 5060 laptop (WSL2), the author runs Flux and Wan 2.2 fully locally. The bottleneck was settings, not the card: Flux fp4 (svdq-fp4r32 via nunchaku) at 6.6GB fits fully in VRAM at 3.4-3.8s/step vs 11-22s/step for the smaller Q4KS GGUF that offloads 1127MB; cpuoffload=auto dropped GPU usage to 1% at 234s/step; and Wan ti2v-5b q5 runs 960x544, 81 frames at 1.95s/step with 3GB free — far beyond the 480x832 the author assumed was the ceiling.
More from Infra
- Anthropic S-1 reveals 48 cents of compute cost per dollar of Q2 revenue — nmasc_ · 2026-10-26
- Persistent AI joins UR network to route private AI inference with encrypted, IP-hidden traffic — 0xSammy · 2026-10-26
- Anthropic's disclosed compute commitments: $114.5B Google, $161.2B Broadcom top the list — BenBajarin · 2026-10-26
- SpaceX strikes $40B deal with Apollo Global to lease AI chips and data center hardware — Polymarket · 2026-10-26
- SpaceX AI revenue up 830%, Claude designs 318K antivenom proteins in 64 hours — PeterDiamandis · 2026-10-26
- Meta plans up to $95 billion stock sale to fund AI buildout — The Decoder · 2026-10-26