2 questions: 1. Can you add in the previous benchmark of requirement to run intern 1 mil context sizes etc?

[sprappcom]

1. Can you add in the previous benchmark of requirement to run intern 1 mil context sizes etc? to understand how much vram and ram is needed for other models.

2. i was wondering if i use a powerful 32gb 5090 gpu paired with dell r930 1.5tb ram. is the ram (ddr3 / ddr4) important in inference speed?
   how fast / slow is the inference speed in this kind of configuration?