Local LLM upgrade dilemma: Dual NVIDIA or switch to AMD?
thatObstinateGuy · reddit · 2026-08-21
User experiencing hallucinations running quantized Qwen3.8 27B on a 16GB RTX 5070 Ti plans to upgrade to 32GB VRAM. Options include adding a second NVIDIA GPU or switching to a 32GB AMD Radeon R9700 to cut costs.
More from Infra
- Pretraining Potential: Coding Agents and the Compute Bottleneck — zeeshanp_ · 2026-08-21
- The Math: Claiming 100T Tokens/Day Would Need ~580K GPUs — teortaxesTex · 2026-08-21
- Moore's Law Fading: Non-Silicon Computing and Novel Architectures to See Capital Influx — MikePFrank · 2026-08-21
- "Why do we need more datacenters? Just write faster kernels" — basedjensen · 2026-08-21
- Cornell Nested Architecture Cuts Training Compute by 36% — burkov · 2026-08-21
- Tencent releases FlashPrefill V2 for efficient long-context LLM serving — tencent · 2026-08-21