This is not the only difference. nForce3 supports AGP (for video only), while nForce4 supports PCI Express (high bandwidth for video cards and other devices). nForce4 also supports SATA2 at 300MB/sec.
-S
PCI clock always rises with the HTT on this board no matter what I try. Overclocks become unstable even with HTT as low as 210MHz as a result. I've verified with ClockGen that PCI is not locked, whether I overclock from the BIOS or from Windows. Any thoughts on how to get past this problem? I'm...
Remember that it is always a good idea to clear the CMOS and redo your settings after updating to a new BIOS revision. This may very well explain the crashing.
-S
The memory settings aren't setup very intuitively in the A8N SLI Deluxe BIOS. All the memory clocks you see listed are based on having your HTT fixed at 200MHz. As soon as you change the HTT, you change the memory clock by the same percentage. As an example... a memory clock of 400MHz really...
Heh... that first pic may as well be the inside of my own system. I have the exact same trio of coolers, although I have a fan mounted over the video card as well.
-S
Hmmm... I wouldn't be so sure this is "illegal". Is it also illegal for AMD to disable higher multipliers on "lower-end" Athlon64 chips? Is it illegal for NVidia to disable Quadro features on GeForceFX chips? The truth is... the law probably allows them to do what they are doing, because they...
http://www.zalmanusa.com/usa/p...ew.asp?idx=71&code=014
Fits perfectly... and even has cutouts in the heastsink so that this can fit on your video card(s):
http://www.zalmanusa.com/usa/p...w.asp?idx=138&code=013
I have both.
-S
You might want to check once more... I find 3 benchmarks where the gain is greater than 5%.
From the article's conclusion:
"The MaXLine III performs just as well as any of the fastest desktop hard drives available today, but when used with an NCQ-enabling controller, the performance...
"Kill" is probably a bit of an overstatement... All the 2x1GB kits I know of are CAS3 at 400MHz, but the actual performance difference in real-world applications is probably less than 1% most of the time.
-S
Alright... $667:
http://www.gameve.com/gve/stor...s.aspx?sku=VC-EVGA-057
...although a heck of a lot of people get by just fine in pretty much all current games with a single one of these... and dual 6600GTs will do even better in Doom3 and the like. How did you survive before SLI existed...
PCI-E 6600GTs are out there in abundance... and are probably the best performance per dollar of any video card available. Beyond that... an SLI 6600GT setup is the -only- way to get the performance it delivers without spending more than $400.
-S
...and no SLI support at all. "Dual" PCI-E isn't very useful until there are cards that take advantage of it. Intel users, at this point, are not using more than one PCI-E slot... because they can't. The only way to SLI is an nForce4 board.
-S
I doubt they care much. PCI-e is neat... and will be a great thing down the road. As of this moment, however, the only cards available are video cards... and those don't really see a performance boost (yet). The one exception is SLI... which is only available on AMD platforms anyway. So maybe...
I'd really love to see someone post benchmark comparisons with 1T vs 2T timings. Compared to known benchmark results for Sandra and all three versions of 3DMark... I seem to be doing great with 2T timings. How much of a gain would 1T be, specifically? Anyone have numbers?
-S
It appears the limit for HTT on this board is somewhere around 340MHz. It's stable up to that point... but things start to get unpredictable after that.
Also... with the memory I have, I can't seem to get 1T memory timings to work reliably. The system boots into Windows, but will inevitably...
Okay folks... I've been messing with this board quite a bit today, and have really pushed it to new limits pretty easily. HTT speeds can go well beyond what we've all seen so far.
My current overclock:
Athlon64 3200+ (90nm) at 2600MHz
HTT at 325MHz x 3 (that's right)
Memory at 520MHz...
Not necessarily. It depends entirely upon what game we are talking about, and how extensive shader implimentation is. A 6600GT SLI setup is significantly faster than a single 6800NU running Doom3 at high resolution. In fact... it's about 15-20% faster than an X800 XT.
-S
After discovering that I had one bad module that only worked single channel (replacing just that one fixed it)...
PQI Turbo PC4000 (TCCD) works great. Running at 500MHz with 2.5-3-3-7 timings.
-S
I don't believe any of the software applications are reporting core voltages correctly. I am seeing the same reported voltage issues as you... but there is definitely a stability difference for me going from 1.5V to 1.525V in the BIOS... even while software shows no change.
I'm pretty sure...
Wow. Amazing the things people complain about.
Back to the point...
I'm having some trouble with PQI PC4000 working dual-channel, so you might try one module at a time as suggested here.
-S
I have a matched pair of PQI PC4000 modules. They use the excellent Samsung TCCD chips. Works great single channel up to and beyond 500MHz (either of the two modules)... but as soon as I add the second, I get a "system failed memory test" beep and the system won't post. I've even tried...
It's probably a good idea in this case to avoid the ones from NVidia. They are all older than the drivers ASUS is providing... and don't claim to support nForce4 at all.
-S
It definitely was not locked in eaerlier BIOS revisions. Using a PCI Firewire card (for more ports) would always result in corrupted FW hard disks whenever my system was overclocked (all other things being 100% stable).
-S
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.