| G.SKILL Ripjaws-Z 16GB DDR3-1600 Memory |  |
| Reviews - Featured Reviews: Memory | ||||||||
| Written by David Ramsey | ||||||||
| Wednesday, 11 January 2012 | ||||||||
Page 4 of 6
AIDA64 Memory TestBenchmark 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 it's a wash in the write test. If you're impressed by the 8.93% better Read score the DDR3-2133 memory (red bar) turns in relative to the Ripjaw DDR3-1600 memory (brown bar), bear in mind that you'll pay four times as much for the former as you will for the latter ($318 vs. $80). The generic DDR3-1333 memory trails in all three tests. The Corsair memory's slightly tighter timings give it a small advantage over the G.SKILL memory in the Read test. SiSoft Sandra Memory TestSiSoft'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. The tighter-timed Corsair 1600 memory turns in fractionally better scores than the tighter G.SKILL 1600 memory, which retails for $10 less. Euler 3DEuler 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 5% over the much cheaper Ripjaws 1600 memory. The G.SKILL and Corsair scores are virtually identical.w Join me in the next section as I run the application benchmarks.
|
||||||||
Comments
I have it in 2133 and 1600 speeds and do not see any real world difference between them.
So the compatibility and low price that GSKill brings to the table have me sold.
"The label shows the XMP profile settings of 9-9-9-234"
Typo I think.
Great review. I used G-Skil 1600 for the first time in my most recent personal build, and they just worked. They drew no attention to themselves which is really all I ask of RAM. I got the black ones and they look pretty sweet, but they're even taller than the quad kit units.
I'm thinking maybe of going Gigabyte for my mobo. If I use G.SKILL DDR3-1600MHz on it, will it automatically give me the same 1.32 vcore value, as well? Must run pretty hot on 1.32v core voltage at idle. (I'm assuming that's the 24/7 voltage.)