bottlenecked

[WxChaser]

I can't figure out exactly why I'm running into a brick wall while benching.

My Abit NF7-S rev 1.2 mobo hits a bottleneck ~200 FSB, with my Barton 2800+ (3rd L12 cut), and whatever memory I throw at it.

I just tried running TwinMos PC3700 in each DIMM slot and benching, and it made zero difference between the 256 stick of TwinMos or the 512 sticks of the Corsair XMS3500!

I have a vdd mod on the mobo, at 1.825v (even adjusted as high as 1.850v with no difference & measured at both points on board). I have my own measuring point as part of the mod with a spade for my positive readout taped to my usb wires so it doesn't move. I've tried higher and lower vcores and it makes no difference also (running normally at 1.7v). Tried 1.875 vcore today and still no diff.

I can boot into windows at 215 to 220 FSB with different multi's but never 3d stable. The memory passes all testing with memtest, doc memory, you name it, test 5 is fine. I've booted up to 450 MHz on the new TwinMos and at least 442 on the XMS3500 sticks.

Can you guys give me a clue here as to what you think it is I'm battling? I don't quite understand. The highest 3dmark 03 score I've gotten was a 6429 yesterday and my settings are all listed below.

6429 (http://service.futuremark.com/compare?2k3=611486)

Thanks for any help you guys can give me. Trying to decide whether to order a 2.0 version board, or just hold off and buy a Canterwood and utilize all this memory! This pisses me off, lol.

I can't keep buying Abit to increase their profit margins when the boards are not quite up to par, if that is what it turns on to be (my biggest suspect at this point). Forgot to add, great cooling on this beast - all air - TT smart fan 2 on CPU and exhaust, 2 80mm intakes, PC Power & Cooing 510W custom PSU, and a Radeon 9800 Pro. The NB has a copper Iceberg with AS3 underneath and a small passive copper sink on the south bridge.

Using a well lapped SLK 800 with AS3 for the Barton.

[SektorX]

perhaps heat on the southbridge is your enemy.
I put a small heatsink on it, just didn't cut it.
Now I have a huge cpu heatsink on it with a fan and it gets me way further

[WxChaser]

When I touch the SB small copper hs...it is just lukewarm though! Even right after priming or running 3D.

I've been looking for a good 5cm fan for it, but maybe I'll look for a larger sink - would probably have to cut my own for it. Did you use thermal epoxy to hold it on on yours?

[SektorX]

well, actually i didn'g attach it just yet.
I just put some as3 on the sb put alot of pressure on the sink and then it just keeps on sitting there pretty stable.
I'll change it soon though. I'm gonna put my homemade waterblock I use on the NB right now onto the southbridge and do something else with the NB. 34°C on NB just isn't good enough :D

[Cucumber]

Perhaps adjust your I/O voltage? As was suggested in this thread....


I/O Voltage (http://www.xtremesystems.org/forums/showthread.php?s=&threadid=11489&highlight=Solution)

...or perhaps even this one:confused: :confused: ...although i swear April fools is at the beggining of the month :p

IDE Cables Causing Problems ??????????? (http://www.xtremesystems.org/forums/showthread.php?s=&threadid=11688)

[CCW]

passive MOSFET cooling could help, they get pretty hot when OC'in, especiually as the proc OCs higher and higher and therefore draws more electrikery ;)

[GVCryan]

It stable in everything but 3d past 200?

I would guess its either drain your 12v line or the raddy not liking the FSB.

[WxChaser]

No - I have a PC Power & Cooling custom made TurboCool 510W with 3 external pots...I never have any drop offs in the voltages either monitoring through MBM 5 or taking my own multimeter readings, or by looking in bios.

I'm pretty sure it's not the Radeon, but who knows? These drivers aren't completely tweaked yet, but I know OPP, Macci, et all are doing well with theirs, although they are pelted & chilled, and mine is just air cooled. I did take off the cheesy radeon HSF/fan combo and reapplied TIM, just doing infinite loops of 3D03
at 10.5x200 now, no artifacts and running great. It's just when I want to squeeze more out of this board it craps out on me.

I'm big time frustrated, but there is always a way.

[WxChaser]

I bit the bullet and bought another frikkin Abit board - revision 2.0.
Should arrive by middle of next week.

If this board craps on me too as far as the >200 FSB bus speeds and my Barton go, it's time for a trip to the dark side again!

I just want 210-230FSB so I can reap some bandwidth benefits, I know my memory is capable of it, either the XMS3500 or the newer TwinMos PC3700 I just got (sitting in a store room getting no attention whatsoever).

I did find a good update on the newer revisioned boards at techseekers.

First though check out this funny animated gif of board changes:

Animated NF7-S boards (http://techseekers.net/images/reviews/nf7s/webdifferences.gif)

Whole review is here:

http://techseekers.net/modules.php?name=Reviews&rop=showcontent&id=48

Wish me luck...I'm going bankrupt here.

[WxChaser]

The dilemma as told by LostCircuits:

Increasing the chipset voltage to 1.7V added stability at high frequency whereas increasing the AGP voltage, once again, decreased system stability. Given everything we just mentioned, it would be easy to blame the board power for the issues we saw, but most likely, the real culprit is something we have seen and documented for the past 5 years and that is the strange phenomenon that high clock speed on the processor end of things works against high FSB frequencies. The only possible explanation is that higher CPU clock rates and the associated performance increase causes higher data demand which, in turn, causes more data traffic which becomes the limiting factor. Granted that this should not happen but we have seen it over and over again and so we assume it does.