Quiz · 5 questions
🎛️ The GPU and Unified Memory
A phone GPU's habits, one pool for two processors, where a discrete card's copy really lands, the boundary that was erased, and the GPU's public doors
Level 0Spec-Sheet Skimmer
0 XP0/91 lessons0/19 achievements
0/100 XP to next level100 XP to go0% complete
Quiz
01On an M3 Ultra, a streaming add over 4 GiB arrays (read two, write one) moved about 635 GB/s against a spec of 819, and a read-only sum 638. What is the correct reading of that gap?
02In PyTorch on a Mac, tensor.to('mps') took 66 ms for 1 GiB. Which statement is true?
03The household's GPU-arbitration stall survived a 2013 dual-GPU Mac Pro, an iMac Pro, a modular Mac Pro and every Ultra. What does the track conclude from that?
04Which of these is NOT a real door to the GPU on Apple silicon?
05What does a well-behaved GPU claimant on a shared-memory Mac do that a process on a discrete-GPU machine never has to?
Comments 0
🔔 Reply notifications (sign in)Sign in — Please sign in to comment.
No comments yet — be the first.