Best chip to overclock

[GTac02]

Ok, which chip will give the best performace, as in higher 3dmark scores, when overclocked to the average overclock that people have been getting.

People have gotten me confused now. Someone with a 2.8c oc'd to 250fsb claims to have better performance than with 2.4c at 300fsb.

 

[goliathvt]

3dMark scores largely depend on memory and graphics card horsepower.

Were the memory modules in these two comps identical? What were the timings? Was a divider used?

What kinda vid card was in each machine?

Remember that 3DMark is just a benchmarking tool. I would be more concerned with real-world performance weighing in on my decision.

As far as which is a better chip to overclock, it all depends on what you want to do and how high you want to get. Unfortunately, there's no way someone can give you a straight answer. However, I will brainstorm a few things to consider:

First, 2.4C's tend to run cooler than 2.8C's, so if you're not into watercooling, that might be one advantage to go w/ a 2.4C, as you don't have to bend over backwards with cooling strategies to get a decent overclock. Also, ZipZoomFly and Azzo are still selling m0 stepping 2.4C's, so there's a good chance you'd get a good one if you ordered now/soon.

On a decent 2.8C with good air cooling, you could probably get a 3.5GHz o/c.

On a decent 2.4C m0 with good air cooling, you could probably get a 3.5GHz o/c.

Another thing to consider is that if you get a 2.8C, you'll probably need to get better memory (PC3500 or better, as PC3200 modules would limit your o/c) for a good o/c than what you'd need to get if you got a 2.4C (they do well with PC3200 and up).

Basically, it all depends on your own goals, cashflow and what components you use to put it all together.

Goliath

 

[johnflyons]

2.8c MO (SL6Z5), Corsair DDR 4000 (520mhz) 1:1, appox 3.6 w/ Abit A17 or Asus p4c800-e

 

[InvisGreenMan]

I just odered one from here at $175. They are supposed to be SL6Z3 M0's. I will post again when it is here. Supposed to be here this afternoon, according to the shipping tracker.

http://www.memorylabs.net/inp42ghz80fs2.html

 

[InvisGreenMan]

Yup, it's a SL6Z3 Costa Rica!

edit: Btw, week 32.

 

[VictorPhung]

Yup, it's a SL6Z3 Costa Rica!

edit: Btw, week 32.

How does your Z3 works InvisGreenMan

 

[Methodx]

3Dmark is almost all about graphics cards.

Consider this:

I originally had a 3Dmark03 score of 1300. When I upgraded from an Athalon 1600 and slow fsb to a Pentium 4 2.6 ghz and 800 fsb, my 3dmark score only went up 800 points!! But when I upgraded from a GeForce 4 ti 4600 to a Radeon 9800pro, my 3dmark03 went up to where it is today.

 

[batboy]

3Dmark2003 is almost entirely vid card related, but 3Dmark2001 is more overall system related with an strong vid card bias. Regardless, both 3Dmark programs are mainly gaming benchmarks and in gaming you want the best vid card you can afford.

Goliathvt had some good points. A properly set up 2.4C with nice RAM with super tight timings that uses BH-5 chips will probably edge out a 2.8C running at the same speeds. But, cranking up to those insane high FSB speeds with a 2.4C is tough if you don't have good cooling and compatible RAM. Often you need to mod the northbridge cooling and sometimes you need to do a vdimm volt mod. That is for hardcore overclockers.

The 2.8C is lots easier to overclock. Sure, you might lose a few benchmarking points, but nothing you'll actually feel in real apps. Most 2.8C will get nice overclocks, usually 3.4 to 3.5 gig on air. You have more RAM options with the 2.8C. You can either use PC4000 RAM with the 1:1 ratio and relaxed timings or PC3200/PC3500 RAM with the 5:4 ratio and tight timings.

 

[InvisGreenMan]

VictorPhung, I get my new motherboard tomorrow. I will post my results as soon as I can.

 

[pkrew]

It really depends on how you set each one up and on the mobo your using. I have a 2.8 that I'm running at 3.46 using fsb 247 1/1 and a 2.4 mo that all I can get is fsb 290 or 3.48 at 5/4. On 2001se the 2.4 is 500 points slower. Every thing is identicle except the chip. The main reason for the descrepancy is the GAT settings that I can use with the 2.8 running at 1/1 and the difference between memory speeds 247 vs 232. The max3 that I have will not allow me to use any of the GAT unless I'm at 1/1. I have run the 2.8 at 247 and with GAT at auto vs F1 and there is a difference of over 400 points or about 2%.

This may not be true for all mobos. The AI7 that I have will allow me to run 5/4 and select speedracer. I haven't run 2001se on the AI7, but it will probably be slower because its a 865.

IMHO this would be the results between the chips using the max3

2.8 using 1/1 vs 2.4 using 3/2 at same speed. 2.8 hands down.

2.8 using 5/4 vs 2.4 using 3/2 at same speed. Toss up.

2.8 using 5/4 vs 2.4 using 5/4 at same speed. 2.4 wins easily

2.8 using 1/1 vs 2.4 using 5/4 at same speed. 2.8 by 500 points!

2.8 using 1/1 at 247 vs 2.4 using 5/4 at 300. I cant get to 300 fsb with my 2.4 but I doubt the extra 10 fsb would make up the 500 points so I'd have to say that with my system the 2.8 would win.

Your mobo and system may vary