The inversion is the part I didn't expect: the dense Zen 6c is now the top IPC core (full 4MB L3 per core), and the old classic design is the high-frequency one. They also tried not to leave a software-visible gap, so the same optimizations still apply to both.
agreed.. Full-size L3 plus more generously sized reordering buffers.
I'm curious if there might be a mixed-CCD desktop part in the future; A 12-core frequency-oriented CCD + 32-core dense CCD would be quite a beast. Any workload that can scale to 16+ cores would almost certainly love to have double the cores at a bit less frequency...
That's essentially what my 7950X3D already is, one CCD is X3D and lower frequency, the other is non-X3D and higher frequency, it works very well.
Realistically the only thing where I see a genuinely noticeable difference between where stuff is running is games, games seem to really love the larger L3 cache.
I think this would be a much bigger deal if it wasn’t for the ai taking all the air out of the room.
Interesting article. Very rare to see CPU designers go so in depth.
>> Very rare to see CPU designers go so in depth.
Very rare to see an inteviewer who has tested to that level and can say "gee you increased L2 hit rate by XX how'd you do that?"
Pour one out for Anand and anandtech.com
The inversion is the part I didn't expect: the dense Zen 6c is now the top IPC core (full 4MB L3 per core), and the old classic design is the high-frequency one. They also tried not to leave a software-visible gap, so the same optimizations still apply to both.
agreed.. Full-size L3 plus more generously sized reordering buffers.
I'm curious if there might be a mixed-CCD desktop part in the future; A 12-core frequency-oriented CCD + 32-core dense CCD would be quite a beast. Any workload that can scale to 16+ cores would almost certainly love to have double the cores at a bit less frequency...
That's essentially what my 7950X3D already is, one CCD is X3D and lower frequency, the other is non-X3D and higher frequency, it works very well.
Realistically the only thing where I see a genuinely noticeable difference between where stuff is running is games, games seem to really love the larger L3 cache.
chasing clocks has a cost.