Dual 7900XTX Owner Weighs Waiting for LPDDR6 Platforms vs Going WRX80
MikeSouto · reddit · 2026-10-07
The user runs a 27B model on dual 7900XTX with a Z390 board (PCIe 3.0 x8) and considers moving to X570 (PCIe 4.0) or X870 (PCIe 5.0) for better performance, while waiting for Medusa/Spark-class platforms with LPDDR6 and 256GB RAM. After seeing Gorgon Halo pricing, they worry about cost and debate going WRX80 with 8-channel memory instead to run larger quantized models (QFN). No plans to add GPUs; Apple is ruled out.
More from Infra
- A DIY micro-cluster: M5 Max 128GB plus Linux box with RTX 6000 for local AI — pcuenq · 2026-10-07
- Running MiMo 2.6 Flash across an RTX 6000 and M5 laptop at 40 tokens/sec over 10 GbE — pcuenq · 2026-10-07
- Running MiMo 2.6 Flash across an RTX 6000 and M5 laptop at 40 tokens/sec over 10 GbE — pcuenq · 2026-10-07
- vLLM v0.31.0 uses CRIU snapshots to restore live TP1 engines without reloading weights — vllm_project · 2026-10-07
- Microsoft Surface event: Laptop Ultra with Nvidia Arm chips expected as Nadella and Jensen Huang take stage — tomwarren · 2026-10-07
- Rumor: chip startup Terafab in talks with all three logic foundries, Samsung ahead — pstAsiatech · 2026-10-07