Apple Says M1 Pro, Max Designs Are Unprecedented
One of the major selling points of the M1 Pro and M1 Max chips is the fact that both chips are system-on-chip designs that incorporate the CPU and GPU on the same die, allowing the two processors to share the same pool of DRAM memory.
“No one has ever applied a system-on-a chip design to a pro system until today,” said Johny Srouji, Apple’s senior vice president of hardware technologies, at the company’s Unleashed event.
Srouji said the SoC designs of the M1 Pro and M1 Max are unprecedented because traditional computers for professional users come with a separate CPU and discrete GPU, both of which have distinct pools of memory.
“But a two-chip architecture requires much more power and cooling,” he said. “It also means the CPU and GPU have separate pools of memory, so they have to copy data back and forth over a slower interface.”
By having a SoC design with unified memory, the CPU and GPU of the M1 Pro and M1 Max can communicate with each other faster than a standalone CPU and discrete GPU can, according to Srouji.
For the M1 Pro, this means up to 200GB/s of memory bandwidth, which he said is nearly three times that of Apple’s original M1 chip. The M1 Max, on the other hand, supports up to 400GB/s of memory bandwidth, which is six times that of the M1.
As scaled-up versions of the original M1 chip, the M1 Pro and M1 Max also have larger memory capacities, with the former supporting up to 32GB and the latter up to 64GB.