Actually, PSU should be fine remembering how much power draw you predicted you should be pulling.
How long did you memtest for? Apparently can take a very long time to for it to produce errors, need to test for more than a couple of hours etc.
Testing each card separately is a good suggestion too.
i memtested for ages… but will be trying it again if im stuck
I dont think one card is knackered… as i can still bench 3dmark for 7240 odd which is where the 6870cf should be no issues.
Dont have any other game installed atm lol so will look at that soon.
Pulled the cards out and will try em one at a time first. Bottom card is first to test … put it in and drivers wouldnt detect etc ( havent had that happen before… always found the single card no dramas befoer when switching a card in and out of cf/sli)
Maybe the vrms/ram heats up too much on a card and it doesnt like it??
As to the psu … i cant see that being an issue as these draw the same (or a shade less) power than the 460s did … but then i ran that on the 1156, not the 1366 which will draw more. But i would expect that to show up whenever you first put the load on the cardsand cpu and ram etc together … and lets face it… we know that a 750 can handle the load.
ill see what the single card trests yield ( disabling cf and gaming still hard locked after 3 rounds(1. maps).
Really giving me the shits … but at the same time … its a good hunt trying to problem solve something it seems no one else has had happen lol - (google has yielded sfa)
EDIT - both cards tested on their own … both crapoped out in bc2 lol.
They both ran fine, then map change, once i spawned and began moving, lasted about 10secs and freeze.
Both individual cards then, i had everest setup … nb/sb stayed at 48-49 degrees for the 20min-30min gaming till crashing. So its definitely not them heating up.
The PSU isnt an issue as the load diff is huge.
So im left with ram ( that memtested fine 2 days ago) or software imo.
fucking tired ( bloody w/e with no good gaming is pissing me big time lol) … i think ill wait till tomorrow to do the next set of tests … but im really running out of ideas/options lol.
hrmm… i may have just missed an important bit.
the fraps for single cards was running around 80-100 ( i have all high but min aa and no HBAO or whatever the fuck its called)
my cf was the same i think … maybe bc2 IS the issue. Will put the CF back together and have a squizz tomorrow
In terms of max stress on the PSU, I still like FurMark + Prime95 at the same time (dunno how furmark handles CF tho). If that gives you no drama I’d definitely be leaning more towards a software problem.
The only other thing that crosses my mind is just try and boost your vtt a bit to see if it helps in game (considering that yours is extremely low for the QPI clock you would be running with a 191 base clock). Your memory may not be at fault and if they are running at their rated spec, they shouldn’t be getting anywhere near hot enough to crash. The QPI could be the link that is breaking considering it handles communication between the CPU and RAM.
A Prime95 blend test should help diagnosing that though. A blend test should push your CPU, QPI and RAM considerably hard. People swear by it, I can’t personally see how it is different to running LinX with max RAM but it’s worth a shot.
I’d still be testing the system on other applications first, like furmark and other games.
yeah i was going to out em under full stress today john.
So i powered on a bit longer last night, and i think i may have the error/fault.
The 5 series had mega issues with bc2 … sky flickering/loading delays and times when it bogged right down. The issue was that the cards would get caught out when ramping up/down … the 2d (low power mode) settings were too low and the graphical demand would catch the card out - hang and then more often than not, freeze. When running one card i seen it momentarily hang heaps but recover.
Now im fairly confident this is the issue with me … as every test ive done to date has proven my hardware is sound. Also, like i told Mac, the load screen and cf flickering etc is down right shocking for me (much like it was for the 4/5 series back at the games start.
It means i need to make a preset, and also adjust the reg
will see how it goes.
EDIT :
Yeah Nip i put that right up thinking the same thing… passed linx/prime but died in bc2… so ive been running stock since.
That would make sense considering your system crashes after the next round starts, when the cards would ramp up again.
Seriously thought AMD would have fixed that issue by now. My friend with the 5970 had that same issue, very annoying but easy to fix.
only issue is making the preset due to ccc2 differing from ccc1 when they used profiles… will have to do some reading.
the 11.4 ( not official yet) drivers have apparently improved the sturreing/screen flashing so im gonna try them first.
Thought with the gazillion bsods a format is in order at any rate, so currently doing that first up.
On a side note, considering the frequency of your formats, are you grabbing the BC2 update files or letting the auto-updater do it? Saves a lotta time if you download the updates from Ausgamer and store em.
if i wasnt getting kicked every 2 minutes due to high ping id at least be able to check if the new drivers fucking work.
fuck this shit.
finally got the chance to test the 11.4 drivers… huzzah … nearly 2hrs of non stop gaming… goddamn it felt good :yes:
Blocks have arrived… as i said, just need to organise fittings.
The blocks came with 1/2" barbs even though he said they were 3/8 
so will need to get extra fittings than i planned.
Have the rotary 90’s … but will need a couple more 3/8" barbs for the blocks. Will just need a 37mm extension from TKR to utilise a 90 on the top card from cpu. Deciding which way to run the loop DRIVE>CPU>GPU>GPU>RAD …or DRIVE>RAD>GPU>GPU>CPU
Its not much in it temp wise as the heat for gpus first … but its certainly preferred to run cpu first. The main thing for me is routing … will the cpu first be as neat as the other way. Will spend some time looking at it.
Cleanest route is the way to go, looking at an ugly route is not worth the initial tiny performance difference until the water temp balances out anyway, regardless of loop order.
i agree with nip.
Go the cleanest route, although the cpu in theory should be first on the loop for the best cpu cooling.
Drive>rad>gpu>gpu>cpu.
By far the cleanest. Takes the cross out of the route at the cpu and means short links all the way to the cpu>drive return link.
The 90s proved to be pointless as if im going for cleanliness… ill go compressions ( barbs and clamps out the front of the gpus will look ordinary imo). So 2 more with 90 adapters for the gpu ( plus the extension to bring the final 90 outside the card), got 2 comps for the cpu, just get 1x45 rotary adapter for the outlet to fit the comps and clean the loop sweep up.
will get the order done next week with luck and then get it done 
That’s what most people think about in regards to loop order (including myself until I looked further into it) but in reality, it makes bugger all difference.
http://www.xtremesystems.org/forums/showthread.php?t=225485
Not even a .5’C difference. You have to think that regardless of loop order, everything that will add heat to the loop is still going to heat up the water regardless of where it is in the loop. That’s how I think about it anyway. If it was an open loop and the water going through the blocks was only ever going to pass through once, then yes, the rad should always be before the cpu. In a closed loop, that doesn’t apply as heat added to the loop after the CPU will eventually be fed into the CPU.
http://www.xtremesystems.org/forums/showthread.php?t=253337
Another interesting thread showing the effect of parallel vs series GPU blocks. Once again the difference is too small to matter and for the savings in fittings alone, series makes more sense to me considering it’s only a ~.5’C difference.
hmm really? Well i guess its true the water will only get too 1 temp in the end, but surely the water goin into the rad is cooler when it exits?
1 would think your rad is cooling that water by more then .5 degree C?