790 corruption issues?

[BenchZowner]

This is BS ...

Man, can you relax and take it easy sometimes ?
Chill out.

If you think somebody's wrong you can easily talk things out and play nice.
No need to use such expressions.

P.S. The FSB holes seem to depend on the Memory Clocking mode ( linked/unlinked ) and your selections.

For example ( hypothetical numbers, didn't take any notes or did many tests ) [ but this theory comes from my experience with the 680i ] 400MHz FSB with the RAM at 1777 won't work, while it works with the ram at 1600 and 1866

[dinos22]

i've given some detailed feedback today to nvidia guys

fingers crossed they get to the bottom of this QUICK

[DaMulta]

This morning I started up to get blue screen blue screen blue screen lock in nvp_dspl.


Then I turned my memory to 1000 and it went away....I thought I lot everything on this install....

http://www.newegg.com/Product/Produc...82E16820145183

Must be bad ram even tho it passes memtest, but it could be a issue with this ram and the ASUS 790i.


I havent been able to run them at 1333 but at 1066 from day one.

[eternal_fantasy]

thanks a lot man!
thats really helpfull... if you have time and enough sanity left ( ) it would be cool if you could maybe map that fsb hole space in more detail.
what about the q6600 then? same fsb holes for it as for the 8500?

Sure I'll try and go into finer FSB for the E8400, with a short ORTHOS run and a screeny if I manage to boot into windows.
The Q6600 I can not tame... one minute it boots and are semi stable, the next minute it refuses to boot... I would leave it to anyone feeling more adventurous to map the Q6600. Good Luck!!

please calm down man... just cause you get different results it doesnt mean his results are BS :P
Please post the detailed settings and bios version your using and what cpu multiplier, maybe there is some detail that improves the stability at different fsb speeds

My results are very repeatable as I have listed all the settings and parts I used. So anyone should be able to replicate my results. OBR's result seems to have P1P2 enabled, while I did not.

The S2E is very unreliable... one moment it boots fine, the next it does not at the same settings...

Oh well I will plough on!

[eternal_fantasy]

i've given some detailed feedback today to nvidia guys

fingers crossed they get to the bottom of this QUICK

So they acknowledge the problem and working on some sort of fix??

Any possibility of sharing that feedback?

[dinos22]

well i complained i had some OS corruptions mainly while RAM testing and they asked for more details

nothing specials just detailed hardware configs and bios settings and bla bla bla
when i get some more info will update
lets hope the fix doesn't bring some different problems or affect speed we enjoy now

[eternal_fantasy]

well i complained i had some OS corruptions mainly while RAM testing and they asked for more details

nothing specials just detailed hardware configs and bios settings and bla bla bla
when i get some more info will update
lets hope the fix doesn't bring some different problems or affect speed we enjoy now

haha I see, looking forward to that update!

[xt0m]

well i complained i had some OS corruptions mainly while RAM testing and they asked for more details

nothing specials just detailed hardware configs and bios settings and bla bla bla
when i get some more info will update
lets hope the fix doesn't bring some different problems or affect speed we enjoy now

Mhh.. When I reported mine to the XFX guys, no one seemed surprised..

[saaya]

The FSB holes seem to depend on the Memory Clocking mode ( linked/unlinked ) and your selections.

For example ( hypothetical numbers, didn't take any notes or did many tests ) [ but this theory comes from my experience with the 680i ] 400MHz FSB with the RAM at 1777 won't work, while it works with the ram at 1600 and 1866

yeah thats what i exprienced as well.
if you select unlinked mode but use a memory ratio which exists in linked mode too, you arent really running unlinked, and it might work fine, while REAL unlinked, using some very odd memory divider will not work.

i've given some detailed feedback today to nvidia guys
fingers crossed they get to the bottom of this QUICK

i talked to two people who then talked to nvidia about this as well, the more sides they hear about this from the better

if you can, hook them up with mike!

i think he had way more corruptions than anybody else here. hes using a professional areca raid controller with 2 120mb/s ssds, so restoring a corrupt windows install takes him 1 minute

he said hes blowing windows installes every couple of minutes and then restoring them so he probably has a good "map" in his head of where the "minefield" is more dense and where its not so bad

and im sure hed be willing to trouble shoot and try different stuff for nvidia, as for him its no big deal to blow a windows install and then restoring it. plus he will be spending the next days home anyways
again, get well soon man

[KTE]

Problems like this are usually not a BIOS problem if nV/MFG hasn't given you a fix or replied to testers with contacts with some future BIOS hope mentioned - that usually always means a hardware problem they know about and are working on a new revision to fix for it. We have no concrete proof for that yet but to me, I would gather it's something along these lines

P1/P2.. hmm.. you guys incl. saaya working in RAM firms would probably know - when DDR2 was first released, the white papers all stated channel A, channel B and labeled one channel P1 and the second one P2 for dual channel. I wonder what P1/P2 here is and if the naming scheme derives off something similar...

[BenchZowner]

It looks like a BIOS issue to me.
But I won't be the one to touch the fire m8

If the solution shall require a new revision of the chip... there will be CHAOS

[tet5uo]

It looks like a BIOS issue to me.
But I won't be the one to touch the fire m8

If the solution shall require a new revision of the chip... there will be CHAOS

Then I'm glad I got an EVGA in case that happens.

I remember when they let me upgrade via RMA my 680i AR revision mobo for an A1 rev. when it was found the AR's didn't overclock quads well at all.

[tekjunkie]

hummm...
what mem speed?
did you try different mem speeds?

1333 QDR linked and synced with the memory,evrything on auto and I got a lard lock when I ran memtest from DOS. Actually my board came damaged








I am now trying to talk to ncix and find out my options. What do you guys think I should do refund or replace???

[slim142]

1333 QDR linked and synced with the memory,evrything on auto and I got a lard lock when I ran memtest from DOS. Actually my board came damaged

I am now trying to talk to ncix and find out my options. What do you guys think I should do refund or replace???

I would replace it, if you are unsatisfied (let say corruption, ram or FSB problems) then refund it.

I want to see new BIOSes from ASUS

[eternal_fantasy]

Here is a little more detailed analyses of the FSB hole on the Striker II Extreme with Intel Core 2 Duo E8400

All of the following will be ran at these configuration and settings;

ASUS Striker II Extreme nForce 790i Ultra
CellShock DDR3-1800
XFX 8600GT XXX Edition
WD RaptorX 10K RPM 150GB
Enermax PRO 82+ 625W
DD Torture Rack

Cooling;
CPU: Arctic Cooling Freezer Pro 7
Motherboard: original heatpipe cooler, reseated with MX-2

BIOS Settings;
CPU multiplier adjusted to maintain as close to 3Ghz as possible.
Memory set UNLINKED at 1800 8-7-6-21 1t P1&P2: AUTO
Voltage: CPU set at 1.35V in BIOS, 1.312 CPU-Z
Memory set at 1.9V
All other settings left on AUTO unless stated otherwise.




Upon loading Windows Vista 64, a memory benchmark is performed.
A short run of ORTHOS BLEND of around 10 minute is performed to establish preliminary stability.
A system files integrity check is performed after the ORTHOS run to check for file corruption.

FSB are checked in 25Mhz increments. Where one FSB works and the next setting does not, further FSB speeds are ran with settings in between to establish the starting point of the FSB hole.

FSB350 - boot - window loaded - stable
http://i225.photobucket.com/albums/d...asy_86/350.jpg

FSB375 - boot - window loaded - stable
http://i225.photobucket.com/albums/d...asy_86/375.jpg

FSB388 - boot - window loaded - stable
http://i225.photobucket.com/albums/d...asy_86/388.jpg

--------------------------------------------
FSB389 - boot - hangs before window loaded
FSB400 - boot - hangs before window loaded
--------------------------------------------
FSB401 - no boot
FSB425 - no boot
FSB450 - no boot
FSB475 - no boot
--------------------------------------------

FSB476 - Boot - window loaded - stable
http://i225.photobucket.com/albums/d...asy_86/476.jpg

FSB500 - boot - window loaded - stable
http://i225.photobucket.com/albums/d...asy_86/500.jpg

FSB525 - boot - window loaded - stable
http://i225.photobucket.com/albums/d...asy_86/525.jpg

FSB550 - boot - window loaded - stable
http://i225.photobucket.com/albums/d...asy_86/550.jpg

FSB575 - boot - window loaded - manual voltage increase needed and didnt get it stable enough for ORTHOS run
http://i225.photobucket.com/albums/d...asy_86/575.jpg

Make from it what you will from these data.

I have tried using different memory settings away from 1800Mhz on FSB that weren't able to boot, but they don't seem to help... for example with FSB at 400, setting memory at 1866 still resulted in hanging at windows starting...

[BenchZowner]

for example with FSB at 400, setting memory at 1866 still resulted in hanging at windows starting...

Sorry for getting ya fooled, the numbers were imaginary
I'll check it out later on, have to get some work done first, and give you some specific settings to check if you can boot at 400MHz FSB with them.

[eternal_fantasy]

Sorry for getting ya fooled, the numbers were imaginary
I'll check it out later on, have to get some work done first, and give you some specific settings to check if you can boot at 400MHz FSB with them.

haha yeah I know, I've tried many different memory settings, but none of them seems to have an effect. the above was just an example.

Anyhow looking forward to your results!

[slim142]

Wow, I cant believe that less than 400mhz FSB (388Mhz) still beats 500,525 and 550Mhz FSB in memory read/writing and latencies with L1 and L2 cache. One thing I hate from NVIDIA chipsets.

Is 400Mhz FSB a hole with Yorkfields? Because I might want to keep mines at 400Mhz FSB since it will still be decently OC @ 3.2Ghz with good latency reading/writing speeds.

[Holmer]

I use the Stiker II Extreme. I got the following error during post (trying to boot at 9x 400 fsb linked, sync) "ntoskrnl is missing or corrupt and re-install copy of this file to root/.../winnt32". Actually I got excatly the same error during post when I had a WD MyBook connected by eSATA to the Jmicron eSATA (everything stock in BIOS and the Jmicron controller enabled and drivers installed). Unplugging the eSATA cable or plugging it to the eSATA port on the case solved the problem and let me boot.
I suppose this is a sign of the data corruption problem or?
I tried to run "sfc /scannow" but did not succeed since it was asking for my Windows XP Pro 64 CD (the DVD drive letter have been changed since Win XP 64 was installed).
Anybody else experiencing the "ntoskrnl is missing or corrupt and re-install copy of this file to root/.../winnt32" error?

Thanks in advance!

[|SiLA|]

no problem yet here

running q6600 @3825 .. 1700 fsb mem linked 850 7 6 5 18 1t 1,9vdimm 1,42nb 1,53vcore

[tekjunkie]

Here is a little more detailed analyses of the FSB hole on the Striker II Extreme with Intel Core 2 Duo E8400

All of the following will be ran at these configuration and settings;

ASUS Striker II Extreme nForce 790i Ultra
CellShock DDR3-1800
XFX 8600GT XXX Edition
WD RaptorX 10K RPM 150GB
Enermax PRO 82+ 625W
DD Torture Rack

Cooling;
CPU: Arctic Cooling Freezer Pro 7
Motherboard: original heatpipe cooler, reseated with MX-2

BIOS Settings;
CPU multiplier adjusted to maintain as close to 3Ghz as possible.
Memory set UNLINKED at 1800 8-7-6-21 1t P1&P2: AUTO
Voltage: CPU set at 1.35V in BIOS, 1.312 CPU-Z
Memory set at 1.9V
All other settings left on AUTO unless stated otherwise.

Upon loading Windows Vista 64, a memory benchmark is performed.
A short run of ORTHOS BLEND of around 10 minute is performed to establish preliminary stability.
A system files integrity check is performed after the ORTHOS run to check for file corruption.

FSB are checked in 25Mhz increments. Where one FSB works and the next setting does not, further FSB speeds are ran with settings in between to establish the starting point of the FSB hole.

FSB350 - boot - window loaded - stable

FSB375 - boot - window loaded - stable

FSB388 - boot - window loaded - stable

--------------------------------------------
FSB389 - boot - hangs before window loaded
FSB400 - boot - hangs before window loaded
--------------------------------------------
FSB401 - no boot
FSB425 - no boot
FSB450 - no boot
FSB475 - no boot
--------------------------------------------

FSB476 - Boot - window loaded - stable

FSB500 - boot - window loaded - stable
FSB525 - boot - window loaded - stable

FSB550 - boot - window loaded - stable


FSB575 - boot - window loaded - manual voltage increase needed and didnt get it stable enough for ORTHOS run

Make from it what you will from these data.

I have tried using different memory settings away from 1800Mhz on FSB that weren't able to boot, but they don't seem to help... for example with FSB at 400, setting memory at 1866 still resulted in hanging at windows starting...

Wow,eternal good work. I think when all the dust settles with my S2E and when and if I get a replacement, maybe i would go the other way, i.e, keep FSB/QDR constant and run memory unliked with various frequencies. I really think that it's a memory divider/SPP issue which is gonna get better as the BIOS matures.

[dinos22]

Wow, I cant believe that less than 400mhz FSB (388Mhz) still beats 500,525 and 550Mhz FSB in memory read/writing and latencies with L1 and L2 cache. One thing I hate from NVIDIA chipsets.

Is 400Mhz FSB a hole with Yorkfields? Because I might want to keep mines at 400Mhz FSB since it will still be decently OC @ 3.2Ghz with good latency reading/writing speeds.

when you enable P1 and P2 it doesn't

he is only testing FSB holes here
not performance

[eternal_fantasy]

when you enable P1 and P2 it doesn't

he is only testing FSB holes here
not performance

Enabling P1/P2 brings a whole new set of problems I just don't have the strength and will power to test through...

[dinos22]

lol

yeah for P1 and P2 i need cold temps for some reason to get high FSB lol

[Aivas47a]

Check out the difference in the Everest memory scores between EVGA 790i and Asus P5E3 Premium. Same cpu on each board (q9450 at 8x450), same ram (2G Cellshocks at 1800, 8-7-7-20, all other settings on auto). Of course this is untweaked and the boards obviously use different auto defaults but still ...

790i:



X48: