How is the Apple M1 going to affect Machine Learning? (Part 1)
Introduction
Apple presented the first-ever Apple silicon desktop/laptop processor based on the Apple A-series processors in the iPad and iPhone lineup. They introduced this new processor along with three new Macs, including the new M1 MacBook Air, M1 MacBook Pro, and M1 Mac mini.
These new M1 Macs are advertising incredible specs using the new M1 processor, CPU speeds up to 3.5x faster. GPU speeds up to 5x faster, and using the new Neural Engine, up to 9x faster machine learning. As for battery life, the MacBook Air is coming advertising 15 hours of wireless web surfing and 18 hours of movie playback (a little caveat they have is movie playback using the Apple TV app). Not sure if that makes a difference, but only time will tell when someone on Youtube tests it out. As for the MacBook Pro, its battery life comes in at 17 hours of wireless web surfing and 20 hours of Apple TV app movie playback. I don’t know about you, but 20 hours is incredible [1]. I’m currently running the 16” i9 MacBook Pro, and I thought the 11-hour battery life of that one was remarkable.
M1 Architecture
The new M1 architecture is made using the latest 5nm technology. It has 1 CPU, 1 GPU, 1 Neural Engine, and DRAM units built into the processor. Since the RAM is built…
