d00m_sayer

• VLLM Prefix caching cannot be used with Qwen 3.5 27b ?

  Posted by d00m_sayer@reddit | LocalLLaMA | View on Reddit | 6 comments

• Dissatisfied with how the RTX PRO 6000 Blackwell is performing during AI inference

  Posted by d00m_sayer@reddit | LocalLLaMA | View on Reddit | 63 comments

• Is there any feasible modification that would allow an RTX 6000 to support an NVLink bridge?

  Posted by d00m_sayer@reddit | LocalLLaMA | View on Reddit | 17 comments

• Question about "./llama-server" prompt caching

  Posted by d00m_sayer@reddit | LocalLLaMA | View on Reddit | 16 comments

• Can PCIE X16 Gen4 SlimSAS 8i x2 Adapters be powered by a second PSU ? or does it need the same PSU that powers the motherboard ?

  Posted by d00m_sayer@reddit | LocalLLaMA | View on Reddit | 12 comments

• Can PCIE X16 Gen4 SlimSAS 8i x2 Adapters be powered by a second PSU ? or does it need the same PSU that powers the motherboard ?

  Posted by d00m_sayer@reddit | buildapc | View on Reddit | 1 comments

• Powering a Rig with Mixed PSUs

  Posted by d00m_sayer@reddit | LocalLLaMA | View on Reddit | 4 comments

• How do I supply power to a PCIe x16 to x8x8 bifurcation card?

  Posted by d00m_sayer@reddit | LocalLLaMA | View on Reddit | 7 comments

• llama.cpp is unusable for real work

  Posted by d00m_sayer@reddit | LocalLLaMA | View on Reddit | 19 comments

• Tesnor Parallelism issues

  Posted by d00m_sayer@reddit | LocalLLaMA | View on Reddit | 3 comments

• How to make DeepSeek-R1-Distill-Llama-70B Model reason?

  Posted by d00m_sayer@reddit | LocalLLaMA | View on Reddit | 2 comments

• Extreme heat in Saudi Arabia leads a gas station's gasoline storage tanks to dramatically burst.

  Posted by d00m_sayer@reddit | collapse | View on Reddit | 1 comments