Question Thoughts on Meteor Lake's hybrid approach vs Apple's unified architecture

[Muhammadshafi]

Examining Intel's Meteor Lake efficiency clusters vs Apple's approach with M3. While Intel separates their big/little cores across chiplets, Apple keeps everything on a unified die with shared cache.


Running some tests on both platforms:

• MLK: i7-14700H (6P+8E+2LP)
• Apple: M3 Pro (6P+6E)

Key differences I'm seeing:

• Thread scheduling latency
• Cache coherency impact
• Power efficiency under mixed loads

Anyone benchmarked both? Particularly interested in real-world performance when juggling mixed workloads vs synthetic tests.

 

[igor_kavinski]

Would you like to share your benchmark scores?