Official LanParty ICFX3200-T2R/G Overclocking Help/Discussion

[Thraxz]

I still disagree. I counter and say it's easy to blame it on the user. My conclusion is not only my own. I know professionals (system builders and certified NEs) who have done extensive testing and concur with me--the 965/975 boards are inferior to the 680SLI boards. I'm talking about people who have utilized phase, peltier, water and air cooling solutions.

 

Consider this: Unless you voltage mod the board you can only reach memory speeds around 1050 max (I reached 1029). There's no active cooling on the nb. Those reaching 3600 are most always using wc and if you notice in the OCDB, mem speeds are most always below 1000. The memory is not unlinked from the FSB so you have a limited amount of dividers. The 680SLI boards do not have this limitation.

 

Now I know some are going to say "mem speed doesn't impair performance." Well, it really depends on what you're running. I do a lot of video encoding where mem performance is more important. I purchased low latency PC9280 to utilize it, not to run it at 1029.

 

Also consider where DFI has positioned/marketed the Infinity. It's NOT a premier board like the LP line. If the overclocking potential of the Infinity is just as good as the 680SLI or even the R600, why is DFI pouring so much engineering and resources into the latter? Why waste the money? And why do you see consistent higher overclocks on the 680 chips across the Internet?

 

I don't want to get into a flame war here, I just simply disagree that the Infinity is in the same league as the Nvidia 600 series or the R600.

 

Yes, I couldn't get my memory to run at spec either on this board... at first.... and now I have it at DDR 1050 4-4-4-12. I could also not break past 350FSB... at first... now I have it at 423 CPU limited.

 

Like all DFIs it takes work and help from other clockers about (Raju's updates were instrumental for me... especially regarding Read Delay) to really make it shine. I've not seen many if ANY p5b's run 3.8 simultaneously with cas4 DDR1050 (let alone 8hrs prime stable.) Now... I've seen a few 680i boards that have and give them credit. I will also say that, yes, they do have a higher "DFI-street stable" FSB than the 975x. However, unless you're using phase/E64 or lower you'll never use that FSB. I know I can't and I'm considered to have a great chip.

 

You mention water cooling with the OCDB... what does that have to do with the board? Not very much. We've already shown that temperature limits the >E6400 chips are well bound by temps before the boards FSB limit.

 

Your point about async mem speeds is valid but I've not run into that issue, personally. ~3.8GHz coincides REALLY nicely with ~DDR1050 the max on both my components.

 

What it really comes down to for me is... the RD600 and the 680i are both $$$ boards right now and I'm certainly not seeing any advantage worth the $100-150 more I'd have to spend over my 975x/g to get it. (I got it for $140 shipped... I'm lucky if I can get an asus/evga 680i or dfi rd600 for less than $250)

 

EDIT: Here's a link to my post about a link to the current 680i chipset boards' straps problem.(What a mouthful). I'm wondering how the DFI's fares.

 

RAJU! WE'VE GOT SOME WORK TO DO!

 

http://www.dfi-street.com/forum/showpost.p...2&postcount=325