r9 fury x for local inference
retained text
Trying to repurpose an old PC, which has i7-6700 cpu, 16gb ddr4 and this r9 fury x. The bandwidth is surprisingly good for a 10 yr old card at 512GB/s, but only 4GB HBM.... Is there any hope here to run a useful model locally?
document fetched 2026-10-24 00:42 · canonical https://reddit.com/r/LocalLLaMA/comments/1xd5uxt/r9_fury_x_for_local_inference