I’m looking at two servers: the IBM X3850 X5 (7143-AC1) (four Intel Xeon E7-4830 processors) and the Dell PowerEdge r715 (two AMD Opteron 6380 processors).
The PowerEdge has a L3 cache size of 16 MB (shared with the L2 cache), base clock speed of 2.5 GHz, and 32 threads total.
The IBM server has 24 MB of Intel Smart Cache, a bus speed of 6.4 GT/s, a base clock speed of 2.13 GHz, and 64 threads total.
However, there is a bit of a grey area.
The PowerEdge server offers less overhead and faster INT performance (as it has half the number of sockets). As an added bonus, the Opteron processor also allows for AVX (advanced vector extensions). That said, the IBM server would be faster in FP operations, meaning that it would have a higher number of FLOPS.
So that means that the PowerEdge has: lower overhead, faster INT, and AVX
While the IBM server has: Faster FP operations, more RAM, more cache, and higher base and top clock speed
Both have the same number of cores installed, support multi-threading, and the same amount of total RAM/memory.
Ultimately, which would be best for OpenMC?
Edit: The IBM server is the one with more RAM