This is strange. If single-core performance is improved, that should automatically translate to multi-core performance, i.e. if each core is 20% faster, logically it would follow that the whole CPU with all cores working would be 20% faster as well. Otherwise, it means it can't maintain that increase with all cores running, which hints at the improvements being due to clock speed vs IPC. IOW, these results seem to indicate the performance improvements are due not to improved architecture, but rather higher clock speed, specifically higher boost speed.
Quote from: xpclient on November 26, 2020, 06:46:07
No Thunderbolt=No AMD. When will they take it seriously? I need it for 10 GbE networking.
I believe this should be available via USB 4, which should be on AMD systems next year.