Quote from: nbt on March 30, 2020, 20:18:45
I believe because both are 16 threads, AMD won't be faster in anything. I was seeing how New Xeon R is somewhat faster than Epyc both 48 threads w/o AVX and w/o changing architecture in 14nm++ still. It means AMD's much Higher L3 catche is not necessarily able to catch up to Intel (the Intel needed 25W more based on TDP which is not a lot in 205W). If intel increases catche in xeon ice lake, it will beat AMD badly they can do nothing.
You remind me of a guy who spams Anandtech claiming that L3 cache is the secret sauce to better performance. It's not quite that simple - they don't just get to "increase the cache" without redesigning most of the architecture around that.
Until Intel can show their 10nm Xeon designs in datacentre-level quantities, we'll have no idea how relevant they are to the market.