Hacker News Reader

Isnt that harder than just making one big chip?

Not sure, was hoping to get some insight here. I know NUMA has many significant challenges but so does adding more cpu/gpu cores and on package memory to compete with current desktop/workstation levels on the ram and gpu sides.

Just seems elegant to do 4 M1's and get 64gb total memory and 4x gpu/cpu and seems hard to scale up a single chip to match current multi chip / discrete gpu systems.

Maybe even dimms for main memory and current on package memory acting just as a very large level of cache.

Four M1s in a NUMA configuration would get absolutely crushed in the high end market outside of niche uses because of the amount of interconnect bus stress. NUMA is most advantageous on server workloads that Apple doesn't target and for which Mac OS is poorly suited.

Reply on news.ycombinator.com