Thanks, much appreciated.Quote:
Originally Posted by BenchZowner
Printable View
Thanks, much appreciated.Quote:
Originally Posted by BenchZowner
Guys,
got a big fat problem over here. Finally got the setup together, was a LOT of work with the dual loops and all but now it's done.
However, I can't get the GFX driver installed. S2E bios 0511 running stock speeds with a QX9770 and 2x2GB OCZ DDR3, Tri-SLI with 3 Asus 8800 Ultra cards, Win Vista Ultimate 64.
Even got an install with integrated SP1, tried every driver from the latest official to every beta and it just doesn't work.
In mid-install I always get a bluescreen saying "nvlddmkm.sys" is missing/corrupted or whatever. Help anyone? :/
I will try to reinstall tomorrow, first running windows update and then using the latest Nvidia drivers (9.64) and NOT the Asus CD.
Hopefully it will work. If anyone else here is running Tri-SLI, what drivers are you using?
Try & find beta drivers & see if that helps..
It depends on if you are using 3 x 8800GTX or 3 x 9800GTX I am using 2 x 9800GTX with BETA driver 174.93 seems to be working fine in Vista x32..Quote:
Originally Posted by jcool
Have a look at my post over at the thread "790 corruption issues?"Quote:
Originally Posted by Tyke
Quote:
Originally Posted by eternal_fantasy
So try going back to previous " uncorrupted" setting and "sfc" may report no corruption again.Quote:
Originally Posted by eternal_fantasy
If you are saying that your windows installation is corrupted based on what sfc reports, look @ eternal's post in the 790i corruption thread. I don't rely on it, am on xp anyways(sfc in xp does not report back anything). I don't reinstall my OS until and unless the os tells me something is corrupted(look at my screen shot in the 790i corruption thread).Quote:
Originally Posted by Tyke
If I get a lockup or BSOD, I try different settings. Fortunately for me, the instability happens rather quickly, unlike a lot of other people who are prime stable for 3+ hours but they are not 3D stable, in a sense that their games lockup or BSODs.
Long story short, don't rely on sfc. Test for yourself, if sfc reports the files are corrupted, go to default speeds and run sfc again. You will find to your amazement that the corruption that sfc reported before is gone.If the file on the disk is corrupted, it's corrupted no matter what speed you run.
Also, on a different note, my settings for 450 FSB were not stable. Prime95 freezes the system after running good for 20 minutes, so I dropped my FSB and upped my multi. I am currently @ 410x10 @ 1.4 Vcore.I will post back the results.
Using 3x 8800 Ultra as I stated... but I haven't tried this driver yet, maybe it'll help.Quote:
Originally Posted by Lu(ky
Yes indeed the SFC errors do disappear when I revert back to stock settings!Quote:
Originally Posted by tekjunkie
Currently testing 444.5 x 9 @1.4125 with 1:1 synced memory.
REQUEST AN INFORMATION CAN NOT RUN MY 4 Gb TWIN3X2048-1800C7DF to 7.7.7.20 1800 mhz the rest settings in AUTO. but only 8.8.8.24 1800 mhz the system crashes in CAS 7 settings, voltage set to v2.0 you have a solution for problem?
thanks
Hey guys,
got it sorted out - one of the Ultras 8the middle one) is defective... plugged them all in one after another, and this one gives the nvkpll... bluescreen when booting. Sucks but it's better than a dead mobo, I guess..
Anyway, VERY impressed with the performance so far, this really is something entirely different from 680/780i.
Hit 19950 3DMark 06 with the 2 Ultras at stock clocks ant the QX9770@400*10 - even running 2T mode (mem at 1600@7-7-7-20 sync).
How much NB voltage does it need to run 1T at 1600? Tried 1,4 but it froze after 10 seconds in windows, lol.
Had exactly the same problem over here.Quote:
Originally Posted by jcool
And when they are watercooled that isn't funny to look for the bad one :(
Had to fill and bleed my loop 3 times before i found the defect card.
Nice you found out, and for now, happy benching :)
Ah well... I know from experience that GTX cards can hold about 6 mins w/o water, while Ultras last about 4 mins. With EK full cover blocks, that is.
Since all I needed to know was whether each card would boot into windows, I didn't need to empty/refill the loop more than once.
By the way, 1600 mem with subtimings on auto SUCKS. I put it back to 1300 and my everest score jumped to 10500 read and 8500 write/copy with a latency around 54ns, while 1600 would give me 7500/8500 and 78ns or something, lol.
Problem is, I really don't know anything about DDR3 tweaking yet (don't even know what sort of chips these 2x2GB OCZ reaper 1333 CL4 have, probably not Micron), therefore I have no clue as to what might be appropriate subtimings.
Will need to play around, but I am happy for now with 20650 3DMark 06 (that's what I got with the Ultras slightly overclocked and stock mem clock, lol)
However, I have a lot of experience in dealing with crappy old 680i... and what I noticed is that tweaking PCIe and HTT to 125/250 like we all used to back then isn't making any difference at all.
Guess the Ultras are finally saturated with 790i.
I haven't seen three-figure FPS in deep freeze before though, let alone with only 2 cards and for like 20 seconds up to 114 :D
I'm curious about this too.:DQuote:
Originally Posted by jcool
nice results thanks
I am having major problems with my Striker II Extreme concerning memory.
It seems that I can not get this motherboard stable at around FSB 1600 (1602/1598) with my 2x2 GB OCZ Platinum EB 1600 (OCZ3P1600EB4GK).
I found settings at stock CPU speed, linked, sync, FSB 1602 that would pass Memtest86+ for 2 passes but testing with 32M SuperPi /mod1.5 XS almost immediately result in error.
My CPU runs fine (SuperPi 32M for now) at 3660 MHz just by setting the multi to 11 and auto FSB clock mode. Running at stock CPU speed with 7.5 multi, linked, sync, FSB 1602 (and 1598) or unlinked with memory at 1333 and fsb 1602/1598 is not possible to get stable. I have tried to tweak all voltage settings, primary and subtimings (always using P1 and P2 on auto and 2T and memory volt at 1.86 V), even changed all the GTL_REF, HTT speed and followed the excellent advice and experiences from you in this and other threads (especially Justifire who has mostly similar hardware to mine). I have tested with both BIOS 0511 and 0601 - same issues.
The CPU seems fine but is the motherboard og the RAM faulty? Tomorrow I can test another set of RAM equal to the ones I have now. My feeling is the MB is the problem.
Please give me some suggestions about what is going on and what to do?
I am having similar problems .. a bit of a nightmare. 1600 seems to a problem, even for my memory which is 1600MHz rated, unless the memory is linked. I am currently running at 1680MHz 7-7-7-20 with 2T timings and a 9X multiplier on the CPU. I have been able to get to 1900MHz at 8-7-7-20 2T using stock voltage with this memory, but the CPU is not stable at these frequencies. The memory is very unstable when running unlinked at any speed. . Also have you tried running the memory in slots 1 & 3 instead of 0 & 2. I think there is still a lot of work to do with this board before I get a system that is fully stable.Quote:
Originally Posted by Holmer
Thanks for responding. My memory are rated to 1600 MHz as well so I blame the motherboard so far - which is a nightnare to get to know. I have not tried the two other banks yet. I tried FSB 1800 linked once with very relaxed timings and froze during post followed by BIOS (0601) resettiging itself to default - guess that was a no go. Strange thing is it seems that I gain more stability (Memtest86+ v 2.01) when setting the HTT speed to 210 at stock CPU speed and 1333 MHz RAM (auto). This was not the case with some 2x1 GB Corsair 7-7-7-20 1600 MHz RAM I had shortly. I will test another set of OCZ.Quote:
Originally Posted by Tyke
Running linked, sync at 1333 FSB results in reboots during SuperPi 32M testing for me?
I'm having similar problems. Anything at 1600 FSB just causes problems. I can run 9 X 1800 and boot, P95 24 stable, but If I run a game, and try to tab over and use my web browser, the system just restarts itself. I'm running a memtest right now and It looks pretty good so far. I cant find what is causing the random restarts. I've tried backing down the multiplier, same result. Anyone know of possible causes for random restarts? If I can get rid of this bug I'll be completely stable.
Same here.
Couldnt boot with memory @ 1600 at all
Booted right away @ 1800
Now have sended them back and get my new kit (same OCZ 1600 EB) in a few hours (the mailman is on his way)
Will report back when i tested them.
Whole my life i have bad experiences with OCZ memory, but damn, they are so much cheaper than other brands.
I stil like GSkill and Corsair best, if this time the OCZ fails, i go with one of them.
I have been following your posts closely so please keep posting your experiences. I reverted back to BIOS 0511 and just tried to run FSB 1800, Multi 7, linked, sync, timings 8-8-8-24 2T, HTT 210: It booted and managed 4 loops of SuperPi before making error. Now I am tweaking overvoltage settings to se if I can get i closer to stable.Quote:
Originally Posted by Justifire
Now at:
CPU: Auto
LLC: Disable
PPL: 1.6
VTT: 1.4
MEM: 1.86
NB: 1.56
SB: 1.6
Rest at Auto
Are you still using a VTT of 1.6? Do you change CPU GTL_REF?
can't make it through 32M SuperPi
I just turned of P1/P2 and it seems my random restarts are gone. I'm going to keep trying to duplicate it and see if restarts come back when I re enable P1/P2
I'm using 1.6 for VTT and PLL. I havnt changed GPU GTL_REF
BTW the above is with my 2nd set of OCZ RAM (see sig). The RAM is not the problem. The reason I can now boot at FSB 1800 (as opposed to my post #267) is the change from BIOS 0601 to 0511.
SuperPI freezes constantly now at any voltages (also the ones "passing" 4 loops initially). Sucs big time.
Yes, VTT 1.6Quote:
Originally Posted by Holmer
Later on i will try lowering this and see if things keep stabile.
Have changed some settings with GTL_REF, but without any gains so switched back to auto.
Vcore 1.48750
PLL 1.6
VTT 1.6
NB 1.7
SB 1.6
LLC enabled
Thanks Justifire, I appreciate it. You are using BIOS 0512 right? What about your memory settings, still at 8-7-7-20 2T unlinked? I guess you are testing with your new kit (I'm afraid the new memory is not changing anything, it did not to me at least). Eager to hear your experiences.
I just memtested my new sticks at stock settings (overclock tuner: Auto, muliti 9, HTT 210, memtimings: 7-6-6-24-2T P1+2: Auto, subtimings all auto. Overvoltage settings all auto except memory at 1.86 and it fails test 4 pass 4 (adress 00027d031e8). I have contacted RyderOCZ for help with settings that should be stable and error-free. I need a baseline, if I am not stable at stock speed (memory even at 1333 MHz) there is no point in overclocking the memory. Honestly I did not test my original kit that long at stock settings but they also needed tweaking with the chipset voltages.
As far as I understand when reading in OCZ forum, memtest errors in test 1-4 are often caused by chipset (e.g. improper voltages) so I will continue tweaking the overvoltage settings. Problem is I suppose the overvoltage settings only apply at a given CPU, FSB and MEM speed (naturally with the same CPU and memory). It is just not fun to tweak settings at stock speed ;-)
Thanks Fusion2002, keep posting. RyderOCZ told me to set P1 and P2 to Auto when under FSB 1900 so that is a good move towards stability.Quote:
Originally Posted by Fusion2002