|
DDR2 Memory Comparison Test |
|
|
|
Written by Djordje Kovacevic
|
|
Wednesday, 21 January 2009 |
|
Page 2 of 3
Kingston HyperX 8500
|
|
When HyperX modules certified on DDR2-1066 and 1200 first appeared,
Kingston was using only Micron D9GKX chips on them, but as time passed
by they have changed that policy. Kingston started to use ELPIDA chips
which diminished popularity of those modules among experienced and
better informed users. On our test modules ELPIDA chips were present
also because, as we expected, it is so hard to find these modules with
Micron chips that are not older more than one year. Despite all, HyperX
showed as best modules on EVERST test, if we overlook their access time
that is for 0.1ns higher than GeIL modules. Even though manufacturer
specified that for this HyperX is necessary to set voltage at 2.2V to
work properly, our samples worked at 2.1V. When it comes to
overclocking we achieved 580MHz which is good result for ELPIDA chips,
but for this achievement we needed to set voltage at 2.26V. Kingston
HyperX 8500 comes with coolers, which is standard for HyperX series.
Cooler itself isn’t good as GeIL’s, but memory was obviously
“cooperating” better with our motherboard.
|
Transcend aXeRam 1066
Transcend also, as Kingston, started using ELPIDA chips only on
DDR-800, but now uses ELPIDA chips also on 1066 modules for aXeRam
series. It’s important to mention that DDR2-1200 series aXeRam still
has only Micron D9GHM chip. Nice gesture form Transcend is that beside
barcode is marked little letter on module that tells us what chip was
used. Here is what every letter stays for:
N – Micron
M – ProMOS
Y – Nanya
P – Powerchip
E – Elpida
Our modules had letter E, which appointed right away that test results
will be similar to Kingston modules because it was the same chip on
both modules. Obviously, Transcend “loosed” a bit secondary timing and
certified modules for voltage set at 2V. This can also be concluded
from table with test results, because Kingston was slightly better than
Transcend. Quality of coolers is the same as for Kingston but memory
was slightly better cooled. Overclock potential is identical to HyperX
and we managed to get 580MHz, but with lower voltage, which also tells
about “loosen” secondary latencies.
|
|
|
|
