Cardyak
Member
- Sep 12, 2018
- 72
- 159
- 106
I propose that Intel would be better served, on a FUTURE node, to develope a cove core that sacrifices no resources for HT and instead, optimized for single thread performance. Instead, just offer a few more Mont clusters. The number of cases where this hurts would be much less than the cases where it helps.
I said as much earlier on in this thread, I fully expect a complete reset on the Cove designs, starting from scratch with a clean slate and a new design largely inherited from the Atom team.
I’d estimate that not only will HT eventually be dropped, but also other ideas from Tremont/Gracemont will be bought over, such as:
- Less execution units per port and instead adding more ports to reduce backend contention
- Decode units organised into multiple clusters
- Pre-decode length caching