POLL - Best for overclocking, 2.4C, 2.6C, 2.8C

[mrman3k]

Which P4 would you pick for overclocking, the 2.4C, 2.6C, or 2.8C. Now as seeing that the memory seems to be the limited factor, would it necessarily be true to pick the 2.8C since you could get a higher clock speed at the max memory/FSB speed?

Any info would be appreciated.

 

[imported_zenwhen]

2.4c for air cooling
2.6c for watercooling
2.8c for phase change

 

[SinfulWeeper]

Ahh... I agree with the 2.8ghz. Running memory at anything slower than 1:1 is a joke. But I hit the wrong cpu speed, I accidently put in for a 2.6

 

[mrman3k]

I was thinking of the 2.8C since it would have a higher clock speed at the same maximum FSB due to the memory.

xzenwhenx, why do you say 2.6C is for water and 2.8C for phase cooling? Is it simply due to the high clock speeds?

 

[oldfart]

Running memory at anything slower than 1:1 is a joke

Running 5:4 is ~ a 2% penalty. Nothing you would even notice. I picked (and have on order) a 2.6C. There are a few pretty experienced overclockers who agree (Pastor Jay, Prometheus, Big Toe). The FSB and mem are pushed too hard on a 2.4C. You can get higher with a 2.6C without that problem. 2.8C is a little too $$ right now.

 

[lookouthere]

Why is most people picking 2.4C?
Since 2.8C can produce more power than the others, but like oldfart said, 2.8C is cost much money.

 

[Tom]

Because people are cheap.

 

[ShoNuff]

I am still trying to decide which is best. Not a lot of data is out yet. I guess it depends on how you define best. Before my 182FSB 2.4B C1 I would have thought that the 2.4C was the way to go. Now, I am shooting for 3.7GHz minimum on air.

We will see...

 

[oldfart]

Now, I am shooting for 3.7GHz minimum on air.

Look @ FSB and mem requirements for 3.7 GHz:

3.7 GHz FSB:
2.4C = 308 FSB
2.6C = 284 FSB
2.8C = 265 FSB
3.0C = 247 FSB

3.7 GHz Ram @ 5:4:
2.4C = DDR492
2.6C = DDR455
2.8C = DDR424
3.0C = DDR395

 

[anomaly]

Id personally pick the 2.6c. 3.5-3.6Ghz on air is asking ALOT, but very possible with a good C-Wood board.

 

[mrman3k]

I think if I were getting one, it would be the 2.8C that way I could run a 260MHz FSB with a 5:4 ratio to get around 3.6 on air.

 

[Y23KC]

I have a 2.4c at 3.2ghz right now. 270fsb with 5:4 mem is pretty damn fast, but the it won't go any higher for clock speed even at 3:2 mem divisor. I'm gonna get a 2.6c and use the 2.4c in a springdale board.

 

[jdogg707]

Just got my 2.6C and without much tweaking have gotten it to 3.25GHZ @ 250FSB on my IC7. The temps look good and it is actually faster in a lot of thing than my 2.53 SL6S2 OC'ed to 3.44GHZ! Will be looking for around 3.5-3.6GHZ out of this thing if my memory can hold out, if not, I may be going to find some new stuff! Would definately reccomend right now.

 

[mrman3k]

Let us know how high you get that thing!

 

[jdogg707]

System:

Intel Pentium 4 2.6GHZ Processor @ 3.38GHZ 1.575V
Geil Ultra Platinum PC3500 DDR @ DCDDR416 (5:4) 2-3-3-6 2.7V
Abit IC7 Motherboard BIOS v. 1.3
ATI AIW Radeon 9700 Pro @ 361/351 Cat. 3.4

Quick Benches:

Sandra Arithmetic CPU Benchmark: 10361, 2914/6521
Sandra CPU Multimedia Benchmark: 15585, 24970
Sandra Memory Benchmark: 5468/5486

Will do more benching later and will continue to update with more info...

 

[Duvie]

Originally posted by: oldfart

Running memory at anything slower than 1:1 is a joke

Running 5:4 is ~ a 2% penalty. Nothing you would even notice. I picked (and have on order) a 2.6C. There are a few pretty experienced overclockers who agree (Pastor Jay, Prometheus, Big Toe). The FSB and mem are pushed too hard on a 2.4C. You can get higher with a 2.6C without that problem. 2.8C is a little too $$ right now.

I agree 100%.....The fact these early mobos are not necessarily hitting high enough fsb with consistency to get the 2.4c's much out of the 3.3ghz range....i think the 2.6c will get you closer to 3.4ghz...the 2.8c with a 265-270fsb which I see a bit more common on good ocing boards and the fact most ppls ram is limiting them will likley get close to those 3.5-3.6ghz (aircooled limits)...