790 corruption issues?

[5010]

Sounds like the same freeze encountered by oh so many others including myself until I balanced my FSB & GTLREF properly.

Adjusting the GTL is not like increasing the voltage to some other part of your PC.. You are simply setting a "REFERENCE" by which signals on your FSB are measured.

My settings which I've stuck with and have not encountered any hangs with for over 1 week are:

1.20 FSBV

&

+25mv GTLs

-

Simply put; The FSB voltage is the maximum setting for a "1" passing between the NB/SPP and the CPU I/O across the FSB and a "0" is always = to 0.4V. GTL REFERENCE Voltage can be thought of as the dividing line between these two, a signal registering as greater than the GTLREF is = to "1" and a signal registering as less than the GTLREF is = to "0". All current intel CPU's use this method of communication with the NB/SPP but with high frequencies and or "ring back" these signals can become confused thus causing the system to identify false "1"s and "0"s which causes "data corruption"(Hence wrong signal/corrupt signal) and system hangs/crashes.

One of the reasons why this is most frequently encountered when using SLI is because of the additional bandwidth being used across the FSB which of course is adding strain to the GTL circuit and increasing the likelyhood of a fault to occur.

Same thing applies to video decoding since a very great deal of data is streaming into and out of the CPU across the FSB.

Also in the CRYSIS game menu's where many users can even hear their computers make something of a ringing/whistling noise due to the increased traffic on the FSB for whatever reason(high FPS) and also older games like "starcraft" since the high-FPS/amount of data passing across the FSB is incredible made up of alot of relitively small "packets"

For the Record: I've found that I can remain 100% stable upto 450FSB(Basically my limit due to BIOS/BOARD) using the same 1.2FSBV & +25mv GTLs

[slim142]

Something to ponder.I replaced my OCZ 2X2 GB sticks with Gskill 2X1 GB sticks. After a system freeze of SPI32M run @ 1800/1800 QDR/RAM, I reboot and this is what I get.



After I went to 0601, my OS never was corrupted with my OCZ 4 GB sticks.
But as soon as I put in my Gskill 2 GB sticks, it happens.I know it does not make any sense, coincidence .

Dont say that.... I have a 2Gb HZ kit on my hands already!

[OBR]

I am playing a week with Striker II without any corruptions and another problems ... BIOS 0511 and 0601 ... it seems be ok ...

[Lu(ky]

I am playing a week with Striker II without any corruptions and another problems ... BIOS 0511 and 0601 ... it seems be ok ...

Well not for me.. I went from using a FSB 1800 2:1 ratio memory with my Q6600 no problems. Now I put in my new QX9650 and blue screens/memory dumps on any OC settings with BIOS 0511. I have not tried 0601 yet, but I would think this should work if it did on my Q6600 with no problems?? Any ideas?

[slim142]

Upgrade to 0601...

[saaya]

But since the sun is now shining in this country called UK, yes I know hard to believe. Gonna take a stroll outside first

hurry up! sun in the uk is so scarce every school and all work should be interrupted as soon as it shines, so people can enjoy the few hours of sun

My fault, I really meant C0 instead of C1
But yeah, they mentioned an Engineering Sample (C0 I believe) and they said it was working really good. Cant recall what they really said but I know it was something positive.

nonono, they made a sarcastic comment in regards to c0 ES
they said they had a great time running it at 1333 999 or something like that, hinting that even at those speeds it wasnt entirely stable

i googled for bad pool caller and this can cause this error it seems:
bad BIOS
bad driver, often for usb devices
bad/unstable memory
bad connections inside the PC (cables and cards)

not helping really, i know

another way to check system integrity...
i guess you could let a large videofile loop constantly in the background?
keep a copy of it on another hdd and then compare the two files every now and then... if they dont match anymore, well then the file that was played back got corrupted.

[tekjunkie]

To test 3D stability, I run 3Dmark 06 in a loop for 6+ hours.

[tekjunkie]

Dont say that.... I have a 2Gb HZ kit on my hands already!

It could be just that my sticks are bad. As I had already posted, my sticks can't do 1T at the rated speed and timings(1600 7-6-6-18 @1.9V). I get a whole bunch of errors in memtest test#5 @ 1T, but 2T passes just fine with rest of the BIOs settings the same.

[eternal_fantasy]

Some update: At settings that caused the use of "sfc /scannow" command to show system files corrupted using ASUS Striker II Extreme fails to boot on the EVGA 790i Ultra motherboard. So basically settings that are stable/POSTs on the EVGA will not cause corruption on the S2E.

Also altering the BIOS in the S2E from unstable to a stable setting, or even changing the CPU(e.g. from QX9650 to E8400) and run again "sfc /scannow, would return a "no corruption" result, right after a "corruption found" with prior "unstable" settings.

Seems that with some unstable settings "sfc /scannow" wrongly finds corrupted files, but actually the files are fine, and simply reverting back to stable settings get the "sfc" to scan and report correctly again...

An intriguing finding, and may not reflect on everyone's hardware.

seems that I have been reinstalling (right after an unsuccessful "sfc" scan) windows when there is no corruption!

[tekjunkie]

seems that I have been reinstalling (right after an unsuccessful "sfc" scan) windows when there is no corruption!

hehe


Anyway, I found out about this myself. Whenever I had a system freeze/hang, I revert back to default speed and windows happily boots and is stable with no freezes. If the files were corrupt, the system will not be stable @ default speed. Anyway, the problem that we are having @ high FSBs might just be due to running "out of spec" which probably will be fixed as the BIOS matures

[Xello]

Can we get any more confirmation on this, that a corrupt installation on overclocked settings is incorrectly and only temporarily showing as corrupt?

That would pretty much set my mind at ease with this board (and i imagine quite a few others') as was looking to get XFX reference.

Someone should update the first page with this info if true, it would save alot of reading!

Also any idea why it would be incorrectly reporting this? I also wondered why Nvidia hadn't figured that out already, as they're supposedly investigating it right now.

[slim142]

hurry up! sun in the uk is so scarce every school and all work should be interrupted as soon as it shines, so people can enjoy the few hours of sun


nonono, they made a sarcastic comment in regards to c0 ES
they said they had a great time running it at 1333 999 or something like that, hinting that even at those speeds it wasnt entirely stable

i googled for bad pool caller and this can cause this error it seems:
bad BIOS
bad driver, often for usb devices
bad/unstable memory
bad connections inside the PC (cables and cards)

not helping really, i know

another way to check system integrity...
i guess you could let a large videofile loop constantly in the background?
keep a copy of it on another hdd and then compare the two files every now and then... if they dont match anymore, well then the file that was played back got corrupted.

So I guess there wont be a C0 soon lol. I have no other option but to get a 7x0i motherboard, I have 2 8800GT to run in SLI and they are becoming older as time goes by!

It could be just that my sticks are bad. As I had already posted, my sticks can't do 1T at the rated speed and timings(1600 7-6-6-18 @1.9V). I get a whole bunch of errors in memtest test#5 @ 1T, but 2T passes just fine with rest of the BIOs settings the same.

I have that problem with my striker extreme too, my g.skills wont run at 1T but with 2T they run perfectly. Some other people with OCZ and Crucial memory run their RAMs at 1T normally. I have no idea how, probably more RAM voltage or more NB voltage? Anyways, I dont think I should really panic about this.

[eternal_fantasy]

Just wanted to mention that the "reversible" corruption only applies when you run "sfc /scannow". I have just corrupted my install for real now, and returning to default does not help with the botched windows installation.

[Broly]

Well not for me.. I went from using a FSB 1800 2:1 ratio memory with my Q6600 no problems. Now I put in my new QX9650 and blue screens/memory dumps on any OC settings with BIOS 0511. I have not tried 0601 yet, but I would think this should work if it did on my Q6600 with no problems?? Any ideas?

I got this problem too. Any help?

[slim142]

I got this problem too. Any help?

Have you guys tried 0601? (From pmp)

[Lu(ky]

I got this problem too. Any help?

Not yet!! Planning on trying the 0601 BIOS, but I see allot of guys with same setup that had 0601 and went back to 0511. So as of right now I can not do anything with this board with my new QX9650 CPU. Kinda is pissing me off a little with all of these memory dumps... I am really thinking of buying a x38/x48 board soon if this is not fixed...

[saaya]

guys, cmd is mostly a chipset thing, i dont think its the memory that doesnt want to run cmd 1t, its most likely the chipset... try more chipset volts

[Mezzro]

Can we get any more confirmation on this, that a corrupt installation on overclocked settings is incorrectly and only temporarily showing as corrupt?

I'm seeing behavior that would support this theory. Specifically, corruptions on POST from the OS that erroneously report various and random system files have been corrupted -- usually after increasing RAM speed and/or tightening timings.

Using the "Last Known Good Configuration" under the OS recovery options, there is no LKGC acknowledged -- the computer will get caught in a dui loop cycling through the recovery options, (F8) menu until I select "LKGC" two or three times, at which point it reboots in the OS without any changes to system files. It does not restore a system state or system files, but the OS is tricked into thinking that there is corruption when there is none. The Event Viewer also shows no evidence of file corruption.

I think most of these reports of data corruption are well intentioned but erroneous reports based on misleading behavior from the OS -- false positives. I'm not stating there is no data corruption in all cases, but I would be willing to bet they are the minority of cases --very small percentage.

The only problem I can confirm with the EVGA 790i Ultra board is its extreme instability with more aggressive overclocks. I can run a Q9650 at 3.60GHz 24/7 without any problem; however, anything greater than that and I get the same blue screen errors, data dumps and false system file corruption messages others are seeing. Despite initial reviews to the contrary in magazines like Maximum PC and CPU, I do not see this board as a good overclocking mobo -- not without some very serious BIOS revisions. So far none of the BIOS releases have come close to addressing the myriad of problems.

I don't think the 790i is ready for prime time. I'm starting to miss my P35 chipset. The X38 chipset is much better, but it's still plagued by random freezes/lockups.

[gcardinal]

sorry to ask a bit off-topic...
I am about to get a new board and was pretty sure about the 790i Ultra pick.. but after reading this topic... is it really that good idea?
Maybe Asus P5E3 Premium will be much better alternative (if we look away from SLI) ?.. CPU/RAM is #1 priority for me.

[saaya]

if you dont oc the board and are willing to spend some time tweaking it, the 790 is a good option. but if you want to push things to the max and have it relatively stable or 100% stable, or you want a working system at high clocks out of the box, then id recommend you a p35/x38/x48 board.

[webwilli]

@gcardinal

i sell my Asus Striker II Extreme and buy an Asus P5E3 Premium. the best decision that i could make. never NVIDIA again.

[saaya]

HA! even you sold your 790 board now?
you were one of the few people who seemed to be really happy with their 790 boards what happened? corruption?

[webwilli]

a few days i was happy, but now it sucks and i´m glad that i sell it...

- f**king FSB-holes
- CPUs need much more VCore and VTT on STRIIEX
and more.

weird: i never had corruption issues

now i´m back to good old INTEL. ok, no SLI but much better OC and no bugs and other sh**.
i wouldn´t believe, that I buyed a NVIDIA chipset. NEVER AGAIN.

[memory]

I spent one whole night just to reinstall the OS and now I am having phobia to clock my memory too high 940mhz and above....I will be getting BSOD for sure And the next moment, my OS corrupted

[slim142]

a few days i was happy, but now it sucks and i´m glad that i sell it...

- f**king FSB-holes
- CPUs need much more VCore and VTT on STRIIEX
and more.

weird: i never had corruption issues

now i´m back to good old INTEL. ok, no SLI but much better OC and no bugs and other sh**.
i wouldn´t believe, that I buyed a NVIDIA chipset. NEVER AGAIN.

You sold your board because of FSB holes and because your CPU didnt like the mobo? Wow... but as long as the user is happy, everybody is happy
Well, I guess you can use CrossFire , So whats the amazing OC you achieved now?