Memory Bandwidth: Sandra Lite 2007
Contents
We decided to include a memory bandwidth benchmark on this review because AMD64 CPUs have an embedded memory controller, so we were curious to see how it would perform against the memory controller located on the chipset we used, Intel 975X.
It is important for you to know that on Athlon 64 depending on the CPU model the memory bus won’t be running at full speed. Athlon 64 FX-62 has its memory bus truly running at 800 MHz (400 MHz x 2) but Athlon 64 X2 5000+ has its memory bus running at 742 MHz (371 MHz x 2).
On the graph below we also include the maximum theoretical performance for DDR2-800 memories running at single channel (6,400 MB/s) and at dual channel (12,800 MB/s). Since our memories on both platforms (Intel and AMD) were running at dual channel, in theory we had to achieve something near 12,800 MB/s. Let’s take a look.
Here we can see how using a memory controller embedded in the CPU is far more efficient than using the memory controller found on the chipset. Intel CPUs achieved a bandwidth 14.63% lower than the maximum rate for DDR2-800 SINGLE channel. As we had dual channel available, this results translates in a usage of only 42.69% of the available memory bandwidth.
Athlon 64 FX-62 achieved a memory transfer rate 10.20% greater than the one achieved by Athlon 64 X2 5000+, 59.61% greater than the one achieved by both Intel Core CPUs and 36.27% greater than the DDR2-800 standard transfer rate. Athlon 64 X2 5000+, which access memory at a lower clock rate, achieved a memory bandwidth 44.84% greater than the one achieved by both Intel Core CPUs.
Even with these great results for AMD, Athlon 64 FX-62 was able to use only 68.13% of the available bandwidth, so it seems that AMD still has some work to do on their memory controller.