Synthetic benchmarks like Sandra and AIDA64 are often the most obvious ways to demonstrate the increased bandwidth using faster memory allows. We will start our testing with Sandra. SiSoftware Sandra (the System ANalyser, Diagnostic and Reporting Assistant) is an information & diagnostic utility that works on either 32-bit or 64-bit versions of Windows XP, Vista, 7, Server 2003/R2, and Server 2008/R2. Sandra takes advantage of multiple CPUs and CPU cores, Hyperthreading, advanced CPU instruction sets like SSE 4.2, NUMA, and 64-bit extensions. Sandra includes benchmarks for system memory as well as CPU, GPU, APU (simultaneous CPU+GPU), and GPGPU benchmarks. However, perhaps Sandra’s greatest strength is that it remains one of the most popular tests used to benchmark memory bandwidth and latency, and we are using the latest 2011 version to perform these benchmarks.
On the bandwidth test, as expected, the Vengeance LP is within percentage points of the standard “Cerulean Blue” Vengeance kit we reviewed previously – which has identical timings.
The higher latency of the Vengeance kits compared to the Kingston HyperX is noticeable in the latency test. The Crucial Ballistix kit remains the fastest in both tests.
In late 2010 Finalwire acquired the software known as EVEREST from Lavalys and changed the name to AIDA64. AIDA64 Extreme Edition is a diagnostic and benchmarking tool that offers comprehensive system monitoring of data such as CPU, motherboard, hard drive, and GPU temperatures and voltages as well as CPU and memory benchmarks and system stress testing utilities that are useful for diagnosing stability issues or ensuring a system overclock is stable.
We are using version 1.80.1450 to measure the memory latency, read, and write speeds.
Again the Vengeance LP kit turns in bandwidth numbers very similar to other 1600MHz DDR3 kits, but falls a little behind in the latency category due to its slightly higher timings.
With the synthetic tests finished, we next turned to our real world tests on the following pages.