|Kingston HyperX Genesis 16GB DDR3-1600|
|Reviews - Featured Reviews: Memory|
|Written by David Ramsey|
|Thursday, 29 December 2011|
Page 4 of 6
AIDA64 Memory Test
Benchmark Reviews has used AIDA64 and its predecessor Lavalys Everest for years, mainly for its processor benchmarks and CPU stress-testing features. But it also includes a "Cache and Memory Benchmark" that performs read, write, and copy bandwidth tests on a system's installed memory.
The Kingston HyperX Genesis 2133MHz memory wins on the Read and Copy tests, but loses out very slightly to the 1600MHz memory in the Write test. If you're impressed by the 7.35% better Read score the DDR3-2133 memory (red bar) turns in relative to the DDR3-1600 memory (blue bar), bear in mind that you'll pay over three and a half times as much for the former as you will for the latter ($318 vs. $90). The generic DDR3-1333 memory trails in all three tests.
SiSoft Sandra Memory Test
SiSoft's Sandra Lite is a free version of SiSoftware's "Sandra" benchmarking utility. Its comprehensive memory benchmark tests report a number of items, but we're interested memory throughput tests. Unlike some other benchmarks that merely perform straight sequential reads and writes, Sandra reads and writes different areas of memory, using integer as well as floating point data, all in SMP mode.
The HyperX 2133 memory wins decisively in these tests. Interestingly the looser-timed Kingston 1600 memory turns in fractionally better scores than the tighter Corsair 1600 memory, which retails for $60 more.
Euler 3D is a Computational Fluid Dynamics (CFD) program that is multi-threaded and computationally very intensive. The benchmark version by Case Labs is built with the Intel FORTRAN compiler and uses 8-byte double precision floating-point math. The test case simulates Mach 0.5 airflow over a NACA 65A004 airfoil section. The benchmark score is the CFD cycle frequency, with higher scores being better.
The Kingston 2133 "wins" again, albeit by only 4% over the much cheaper Kingston 1600 memory.
Join me in the next section as I run the application benchmarks.