>>105986851
Because the RTX -90 tier of GPUs isn't restricted by a laptop's area & efficiency requirements.
My point is true the lower the tier you go; it used to be the opposite in the good ole days where a GTX 1060 easily was within 10% of the 980's performance.

>>105986813
11.333 ns first word latency isn't the end of the world; you make tradeoffs by going with higher capacity modules.
Workstation PCs prioritize memory bandwidth over latency.

>why would you recommend 2x32GB
Relic of the old OP, you're the first to cry about this after many different OPs had that recommendation.