Thinking about CPU architecture... (after looking to see if I make my next machines another Xeon-D or and E5 system, or a c3000 if they ever arrive)
C3000 2mb/core (Large L2 only)
Xeon-D 1.5mb/core
E5 2.5mb/core mostly
L3 is used a lot when doing multi cpu, it has an impact on some specific scientific workloads.
What is the effect in normal workloads ? Is Xeon-D disadvantaged by the small cache really in any way ?
Let's say comparing and e5-2620 v4 to d-1541 do we see any performance difference, if I look at a basic cpu benchmark like passmark they are essentially identical are they not ? (Couldn't find AES benchmarks but I assume the e5 is better but just a guess)
(Of course if you need more than 128gb ram then e5, upgrade to dual then e5, if you want 10G network Xeon-D, 45w tdp vs 85w tdp, as long as you live within the limits is Xeon-D then simply not better due to 10G and lower tdp?
C3000 2mb/core (Large L2 only)
Xeon-D 1.5mb/core
E5 2.5mb/core mostly
L3 is used a lot when doing multi cpu, it has an impact on some specific scientific workloads.
What is the effect in normal workloads ? Is Xeon-D disadvantaged by the small cache really in any way ?
Let's say comparing and e5-2620 v4 to d-1541 do we see any performance difference, if I look at a basic cpu benchmark like passmark they are essentially identical are they not ? (Couldn't find AES benchmarks but I assume the e5 is better but just a guess)
(Of course if you need more than 128gb ram then e5, upgrade to dual then e5, if you want 10G network Xeon-D, 45w tdp vs 85w tdp, as long as you live within the limits is Xeon-D then simply not better due to 10G and lower tdp?