Benchmark Results: With the 2P server running a pair of X5460 Harpertown 3.16GHz LGA 771 120W TDP processors and 667MHz FB-DIMMs it is no big surprise that it performed at the back of the pack when it came to memory performance. The pair of Intel Xeon E5520 Gainestown processors operate at 2.26GHz with a full 8MB of shared L3 cache, but a lower QPI of just 5.86GT/s. Thanks to the lower QPI speed the DDR memory can only run at 1066MHz and as a result the memory performance suffers compared to the pair of Intel Xeon X5570 processors running at 2.93GHz with DDR3 1333MHz memory modules. The Intel Xeon X5570 processors also benefit from having a full QPI link speed of 6.4GT/s, which is how they support 1333MHz DDR3 memory modules. The extra QPI link speed, clock frequency and memory frequency make the pair of Intel Xeon X5570 processors stand out from the others rather significantly.
Benchmark Results: When it comes memory latency the pair of Intel Xeon X5570 doesn’t dissapoint and performs much faster than the other two configurations that I was able to run some benchmarks on.
This CPU Queen benchmark is simple integer benchmark focuses on the branch prediction capabilities and the misprediction penalties of the CPU. It finds the solutions for the classic “Queens problem” on a 10 by 10 sized chessboard (see the source of the benchmark
). The new Nehalem based Intel Xeon 5500 series processors do great on this benchmark as well as the Worxx CPU benchmark. Some strange results were observed in the AES benchmark though. The AES benchmark is an integer benchmark that measures CPU performance using AES (a.k.a. Rijndael) data encryption. It utilizes Vincent Rijmen, Antoon Bosselaers and Paulo Barreto’s public domain C code in ECB mode (check the code out here
). It seems that Harpertown does a great job with AES data encryption and was actually faster than the pair of E5520 Gainestown processors.