Now, let's see what we gain on the Productivity and
Gaming side of things.
|
Benchmarks
/ CAS2 Versus CAS3 Continued |
It
does make a difference but is it
enough for you? |
|
Let's
work. Here are our Ziff Davis Content Creation
Winstone Scores at both CAS2 and CAS3 settings with
a 133MHz. FSB and a 933MHz. Pentium 3 Processor.
On the
"light duty" business approach, a CAS
setting of 3 doesn't buy you much. Perhaps
that is because many of the tests in CC Winstone
2000 are not heavily memory intensive and are more
resident in local processor memory.
Now,
we'll take a look at how those setting affect a more
memory strenuous application, like a Quake3 Arena.
Here we
see a modest performance gain of about 6% at the
lower resolution (less fill rate dependant) and
about 3% in the higher resolution of 1024X768.
Either way, a few frames per second extra doesn't
hurt. So, in the end analysis, YOU need to
decide if the performance gain is reason enough to
justify the higher cost of CAS2 PC133 memory.
Finally,
an area that is more compelling in terms of
justification for the higher price tags of CAS2
PC133 memory, stability at higher bus speeds, is the
name of the game with these products. We fired up a
gauntlet "Torture Test" of sorts that was
sure to stress the memory to its fullest in terms of
measuring stability. We also used an unlocked
engineering sample Intel P3-933 processor for our
testing. This allowed us to bump the bus
speeds up and the clock ratios down in order to keep
stability well within spec of the processor thus
eliminating any random crashes due to processor
over-clock.
Here are
our findings on stability for these two modules.
We ran
this Prime95/Quake3 Multi-Task test for over 20
minutes at each bus speed. For those of you who are
not familiar, Prime95 is a very processor intensive
analysis tool, that has shown itself capable of
bringing even the slightest unstable system to its
knees. We then ran an endless loop of Q3 Time
Demos on top of the Prime95 instance, just to thrash
the hard disk and swap information in and out of
memory at a very intense rate. The results are
fairly clear. Both of these modules performed
WELL beyond their specified limits.
The
Mushkin stick was able to withstand a full 155MHz.
FSB setting at CAS2 without one lock up or crash in
our test. The Corsair module booted Windows 98 fine
and ran the test at 155MHz. FSB with a CAS2 setting
but then crashed after a few minutes. Both
modules could handle a 160MHz. FSB but neither would
boot the system at 160/CAS2.
In
short, the Mushkin memory looks slightly more
robust. However, this test could be showing
the margin of error possible due to motherboard
instability with a specific module at these high
clock speeds. It is difficult for anyone to
conclude that at crash at this speed is either the
memory or the motherboard's fault. After all,
at 155MHz. Front Side Bus, there are lot of things
in the system that are running beyond spec.
Our opinion is that both modules performed admirably
but perhaps the Mushkin module is a little more
forgiving.
So, what
can be taken away from our little analysis
here. First, a CAS Latency of 2 setting can
bring a measurable performance gain to high end
systems of today. But you already were
probably aware of that. Second, both Corsair
and Mushkin are putting out VERY high quality
product when it comes to PC133 memory, especially in
the CAS2 camp. So, which of these modules is
for you? Well, that depends on a number of
factors including price and availability of the
modules in the retail channels. Frankly, the
modules were pretty much of a "draw" in
terms of our "personal" competitive
analysis. You are in luck however, the folks
at Outside
Loop have both brands in stock and are ready for
the picking.
For now
we'll give both the Mushkin and the Corsair PC133
CAS2 Modules a HotHardware Heat Meter Rating of...
You
want some action? Get into the H.H. Conference
Room Now!
|