Y
Hacker News
new
|
ask
|
show
|
jobs
by
sgillen
290 days ago
It has a unified memory architecture, so the 128G is shared directly between CPU and GPU. Though it's slower than dGPU VRAM.