Overclocking the
Pentium 4 2.4C : We had a few hours alone with a yet-to-be
officially released Pentium 4 2.4C and a new ABIT IC7 Canterwood
mainboard. We wanted to quickly share what is to come.
For those of you that
do not know, the Pentium 4 "C" processors are the ones that use the
new 800MHz Front Side Bus supported by the recently introduced i875
chipset codenamed Canterwood. These chips also support
HyperThreading technology. You can see the 800MHz system bus and
Canterwood in action in
our recent article showcasing it.
Currently, the only CPU
officially released by Intel that supports the 800MHz FSB and HT is
the 3GHz model or the "3C". It is fairly common knowledge in the
industry that 2.4C, 2.6C, and 2.8C processors are to follow here
very shortly.
What is interesting is that long-time [H]'er,
Dustin Crabtree, dropped me an email and told me that he had bought
a new system that contained a Pentium 4 2.4C. Luckily, he lives very
close by and volunteered to bring his box over for a little OC
outing to satisfy our interests about the impending 2.4C.
Dustin bought the CPU from an online
e-tailer charging a bounty for having the "special" CPU. As you can
see from Dustin's new rig, he is not new to the dark arts of getting
more for your money when it comes to computers. You will be able to
tell shortly from his system choices that he made some wise picks.
Dustin did a hell of a nice cabling job as well. But on to the meat
of the matter.
The P4-2.4C
As we have verbalized
in the past, we are hoping that this 2.4C would be a chip to deliver
stellar OC performance as it is of a new stepping while being a
throwback when it comes to MHz. Quite frankly we are hoping it will
be joining the ranks of the Celeron 300A and the P4-1.6A Northwood.
Dustin's system
was equipped with stock cooling. He had reported to us that he had
successfully played 4 hours of Medal of Honor the night before with
the system set to 275MHz Front Side Bus.
We got Dustin's box
up onto the test bench and hooked up to the Koolance Exos System to give us a bit better
cooling.
Quite simply I think many of us will find
the current enthusiast Holy Grail to be 3GHz. Even better, reaching
3GHz with a 1GHz Front Side Bus...
3GHz
Our 2.4C would do 3GHz at default voltage
and stock cooling as verified by Dustin. With the Koolance system it
worked at 100% stability for the short time we had with
it.
With the 5:4 DDRRam MHz ratio, you can run full DDR400 in
the dual channels supplied by the Canterwood chipset. If you don’t
have the Ram to reach these speeds, the ABIT board also has a 3:2
ratio that you can pick in the BIOS. This would effectively give you
a DCDDR333 memory bus, so PC2700 would do the trick.
The
next logical stop seemed to be 275MHz FSB for some reason. Here we
either have a DCDDR440MHz memory bus or a DCDDR366 memory bus.
Seeing as how this may leave you too much room or simply not enough
with your Ram, I see this not being too popular for guys having to
live on the edge.
3.3GHz
Our 2.4C once again passed with flying
colors under load at default voltage with stock cooling as verified
by Dustin. Surely though there was more left in this CPU, and that's
what we wanted to find out.
3.5GHz
Once again our 2.4C was happy at our
chosen speed of 3.5GHz. This was with a bit extra VCore at 1.7v. To
be honest, this was where we had worked our way down to from 3.6GHz
(More on that in a second.).
To say the CPU was stable here,
even with the Koolance Exos water cooling system, would not be
totally true. We could load both sides of the HT CPU by running two
instances of Prime95 to 100%. It would run through about 8 to 9
tests before one virtual CPU would crash the test but not the other.
Still though we decided to see how it would handle some AGP
action while all of that was going on. We were able to run a full
loop of 3DMark 2001 SE with both instances of Prime95 without
crashing anything. So as you see, the CPU was not stable, but it was
far from "unstable". Certainly somewhere in between.
3.6GHz
We could not come this far and not go for
that next milestone. Sadly enough, with our "meager" water cooling,
we would crash booting into WinXP every time at the 1.2GHz FSB
point. I have no doubt that some heavy duty cooling would get you
there though.
Conclusions & Delusions
To sum
it all up, this one retail Pentium 4 2.4C has certainly gotten us
excited about OCing Intel CPUs again. There is no doubt that Intel
has been fighting for more and more of the enthusiast market share
and this is one more landmine for AMD loyalists to sidestep.
We have the feeling that the OCing success stories will not
stop with Dustin's CPU. I personally think that most of the 2.4Cs
that hit the shelves will have the ability to reach the 3GHz mark
pumped with a 1GHz bus and DCDDR400. Mark my words, this box will be
the backbone of the ultimate budget DOOM]|[ machine you will want
later this year...unless we see really cheap Prescott Pentiums soon.
The folks at Intel want the enthusiasts and their dollars
back. Most of all they want the enthusiasts telling everyone else in
the world that comes to them just how great their Intel powered box
is running. That enthusiast market has been nearly all-AMD till the
Pentium 4 1.6A hit the streets with huge OC potential and Intel
pulled a bit more market back their way. We think this P4-2.4C is
going to do the same exact thing. The market will always be divided
from here out as long as there are competitors, but certainly it
will not be as one-sided as in the recent past.
The next big
piece of the puzzle is what is this CPU going to cost, and I think
we will be seeing prices of around US$180 in the U.S. From what I
understand, there is supposed to be a fairly good supply, but if
there will be enough to fulfill power hungry enthusiasts is yet to
be seen. Of course with Intel putting 3.3GHz CPUs in 2.4C boxes, it
may be tough to keep them in stock for quite a while.
In
closing, let me state that we have looked at only one CPU and it is
very possible that we simply got hold of some good early silicon.
That is not however what we think. Still enthusiasts know that all
silicon is different and can yield different results. At this point
though, I have to think Intel has got the next big thing for
OCers in their 2.4C processors.
![[H]ard|OCP](2.4c-overclock_files/logo.gif){kind=logo}
![[H]ard Menu](2.4c-overclock_files/mainmenu_top.gif){kind=small}
Main Page
Print this article
Email this article
