By Hartley Charlton
According to a plausible new theory, Apple's M3 Ultra chip could be designed as a separate, standalone chip rather than being made up of two M3 Max dies.
The theory comes from Vadim Yuriev of Max Tech, who outlined his thoughts in a post on X earlier today. Referring to a post by @techanalye1 saying that the M3 Max chip no longer has an UltraFusion interconnect, Yuryev suggested that the yet-to-be-released “M3 Ultra” chip would not be able to contain two Max chips in a single package. This means that the M3 Ultra will likely be a standalone chip for the first time.
This will allow Apple to make special tweaks to the M3 Ultra to make it more suitable for intensive workflows. For example, the company could ditch the efficient cores entirely in favor of a high-performance core design, as well as add even more GPU cores. At a minimum, one M3 Ultra chip designed this way will almost certainly provide better performance scaling than the M2 Ultra compared to M2 Max, since there will be no more efficiency loss when using the UltraFusion interconnect.
In addition, Yuryev suggested that the M3 Ultra could have its own UltraFusion interconnect, allowing two M3 Ultra dies to be combined in a single package to double the performance of the hypothetical chip.” M3 Extreme”. This will provide superior performance scaling compared to a four-die M3 Max package and open up the possibility of even more unified memory.
Little is currently known about the M3 Ultra chip, but a report in January suggested that it was. will be manufactured using TSMC's N3E node, similar to the A18 chip expected to debut in the iPhone 16 lineup later this year. This means this will be Apple's first N3E chip. The M3 Ultra is rumored to launch as an updated Mac Studio model in mid-2024.
Tag: Apple Silicon Guide[ 136 comments ]
