众所周知，Nehalem-EP也采用了直联架构，而Penryn还是使用FSB传输的方式，似乎前者的数据看起来似乎更值得我们期待。不过，事实真的是这样吗？
SiSoftware Sandra Pro Business 2009 | |||
测试对象 | 华硕TS700-E6服务器 双路Intel Gainestown Xeon E5506 2.13GHz | DELL PE2900 III 双路Intel Harptown Xeon E5430 2.66GHz | |
Memory Bandwidth Benchmark 内存带宽测试 | |||
Int Buff'd iSSE2 Memory Bandwidth | 20.02GB/s | 6.13GB/s | |
Int Buff'd iSSE2 Memory Bandwidth vs SPEED | 9.43MB/s/MHz | ||
Float Buff'd iSSE2 Memory Bandwidth | 20.02GB/s | 6.13GB/s | |
Float Buff'd iSSE2 Memory Bandwidth vs SPEED | 9.43MB/s/MHz | ||
Memory Latency Benchmark 内存延迟测试 | |||
Memory(Random Access) Latency (越小越好) | 108ns | 108ns | |
Memory(Random Access) Latency vs SPEED (越小越好) | 0.16ns/MHz | ||
Speed Factor (越小越好) | 55.30 | 95.20 | |
Internal Data Cache | 4clocks | 3clocks | |
L2 On-board Cache | 10clocks | 18clocks | |
L3 On-board Cache | 50clocks | ||
Cache and Memory Benchmark 缓存及内存测试 | |||
Cache/Memory Bandwidth | 80.69GB/s | 68.88GB/s | |
Cache/Memory Bandwidth vs SPEED | 38.74MB/s/MHz | 26.52MB/s/MHz | |
Speed Factor (越小越好) | 32.40 | 111.90 | |
| Internal Data Cache | 320.09GB/s | 421.23GB/s | |
| L2 On-board Cache | 291.36GB/s | 122.68GB/s | |
和上一页类似，采用了直联架构之后，Nehalem-EP的内存带宽比原来提升了3倍，这也使得部分项目中得分有所提高。
