Vic

7B indicates a SATA AHCI issue.. or something in the SATA ports out of sorts with the BIOS

Its a very common error with more than just one possible problem but before you start tearing whats left of your hair out I dont think this is a clocking problem but something that goes back to the issue of you not being able to run AHCI/IDE ENHANCED

What it is I dont know but I will see if I can come up with anything

Go into your BIOS and verify something has not changed in the way of the SATA drives... check what they are set to run on ACHI, ENHANCED, ETC.. verify the controller is enabled and verify the boot sequence

If that nets nothing, unplug the tower and verify the power connections and SATA conntections to the drives.. you may even want to swap out the SATA cable for the boot drive.

Power up and try again

if it still fails, try moving the boot drive to abother SATA port.. when you do this you may need to enter the BIOS and reverify everything



This error is very specific to the hard drive controller 'missing' although a high memory or unstable memory or cpu clock can cause anything to pop up the error you are seeing is directly hard coded to a missing IDE or SATA contoller on the buss

vgbaron wrote on Jan 11^th, 2009 at 3:27pm:



there is something very wrong..  CHKDSK C: /r should have run if the storage system is being recognized by the BIOS

If all else fails above..   clear the CMOS on the system with it unplugged, enter the BIOS, exit menu.. LOAD SETUP DEFAULTS   F-10 save and exit    return to the BIOS and reset it back up (no clocking) ... locate the drives and verify they are set correctly

Note.. I forgot to tell you that moving the drive to another SATA port may clear the 7B error but Windows may, or may not boot. The idea is to get this 7B error clear and figure out if this is a bad cable, port, BIOS, or what

twoposts up - resolvd main issue.

should the 450x9 clock work w/o mods in XP64 or might some tweaking need to be done?

First reboot into corrected 450x9 profile loaded XP64 but fairly quickly went o BSOD with System_Service_Exception.

Rebooted again and so far, all seems well.

I thought we had the 450x9 clock numbers already worked out.

Are you still running all of those including GTL?

vgbaron wrote on Jan 11^th, 2009 at 3:50pm:



It could have been something in ACPI resources that needed to clear from switching from IDE to AHCI

the storage controller may very well share with the video adapter in ACPI IRQ assignment

vgbaron wrote on Jan 11^th, 2009 at 3:45pm:



No... but you would probably sound like me when I tried to explain to a judge many years ago how the brunette grinch stole my life

64 bit 174.74 wouldn't install - doesn't support gt280?  Said couldn't find supported hardware,

Trying 181.20's

memset is the same but everest is under 10 grand and over 54 - which is different. But the earlier tests were with xp32/IDE

The above is at 400x9 - 3dMark 14917 - no problems with 180.48 or 181.20 drivers

At 450x9 3dMark locked up with 180.48 and 181.20 drivers

Just to clarify - clock was stable with 0901 XP32 - with the HDD/shutdown issues I flashed to 1003 but never verified a stable clock.  Don't know if that would make a difference.

edit: at 450 OCCT gave BSOD after 10 minutes - running at 400 if it passes I'll kick up he NB and try 450 again.

rgr - that was my next thought. Seems ok at 400 1T SR enabled - OCCt been running fine 30 minutes.

I'll set 450x9 - 2T  SR Disabled and test

make sure DRAM voltage is 2.0

Stable for 1 hour however when running 3DMark - had to abort - graphics were totally munged

Monitor makes clicking noise when changing screens in 3DM

upped NB - same problem however now having difficulty booting into xp again.  Logo show, then desktop, then a garbled windows startup sound and the BSOD - system exception.

doing this at 450 right?

PSU is brand bew Corsair 1000w

Seems like graphics screw up in 3DM and the system BSOD's for a while and then is ok.

Problem with 3D mode????

All XP64 updates applied

remove GTL's got partway thru 3DM but frames wer 7 where usually 60 - then it locked with garbage colors. Will reset GTL's with NB GTL at auto

edit: nb auto - system exception with 3dM plus monitor clicking when switch to 3d.

Vic

I just dont get it

Everything in GTL's both CPU and NB should be correct along with VTT and NB voltage @ 1.55 also seems right

SPREAD SPECTRUMS in the BIOS also must be disabled too but I can not see those causeing 3D glitches

If OCCT is remaining stable and this it only tied to 3D something strange is going on. I know stability in 3D can be borked during use (like crashes in FSX but it flys for a  few mintes) but the errors you are getting are really strange

Do you have your old video card still?

The only other thing I can think of is to get the system back to the original stable settings in OCCT.. all the GTLs, Vcore, NB voltage 1.55, everything as we know stable with OCCT @ 450.. verify

Once stable.... hmmmm...    try DISABLE the memory remap feature first... if still unstable, then I have to suggest uninstalling the drivers and pulling the card completely out.. put your old card in and try that one

if you still see the same exact problem then there is something wrong in hardware elsewhere... if it runs clean then uninstall the video driver and put the new card back in and try it again on the same exact settings.

If it borks then you either have a card that does not like this clock or a driver that does not play well with a clock.. you may need to take a look at Nvidia for older 280 drivers and try a few

If still no go and this all works fine @ 1600 and it worked fine at 1800 in x32 I really have nothing else to add other than a BIOS problem. 0801 worked great for you in the past

On a hunch - I dropped back to 400x9 and redownloaded the DX9.0c update. Ran 3DM and still got a few clicks from the monitor but it completed without problem and the score was 18520 as opposed to 14917 before the update. Framerates were in the 100's as opposed to 60's.

Raised clock back to 450x9 - 3DM locked up.

I'll try your other suggestions and see what happens.

flashed down to 0901 - same problem - does NOT like 3d at the 450 clock.

400 clock - all seems fine.

The card is clocked by the mfgr - I made no changes.

The clicking from the monitor is puzzling - I have never heard that before.

If it stays stable at 3.6 I'll bite the bulklket and leave it alone unless I get a brilliant idea. Don't want to throw more $$ at the problem. Might as well save the cash for the I7's down the road.

IIRC, you mentioned that the first releases of XP64 that had sp2 applied after the fact had 3D issues. This was a MS disk with sp2 on it already but the similarity bothers me.

If I get creative in the next few days, I may pop the 280 out and the 8800 in and see if the problem continues. Too pooped to do that tonight.

vgbaron wrote on Jan 11^th, 2009 at 11:07pm:



alright

I will think it over and post if I get any bright ideas but this just does not make sense.. x64 is more stable than x32, always has been

Im really at a loss

I do know there were a few of these boards reported to have a defect where the heatsink rail lifted off the NB due to a cap placement

http://forums.overclockers.co.uk/showpost.php?s=7bf8dd90e51008da8688580afd14cba1...



memory goes in the white slots and I am sure that is where it is..

It just does not ring right.. the voltage are right and with the exception of NB you should not need to set anything different

I have never seen a Q9650 need a VTT of more than 1.30, 1.28 is the norm but I have seen a few clockers claim they had to raise it to 1.35. If you did that, the 4 CPU GTL's would need to be -43.199mv (rounded to the highest value the BIOS allows, probably -45mv

If there was no change I just can not see that being the target area of the problem


Im just about farting in the wind at this point

lol

yup - pll is 1.51106

going to call it a night and unwind. getting brain cramps   

Make sure your PL is not too low too


and to be quite honest..  if you were not seeing any of this in x32, and, you are sure you were set up with no clock and on the right BIOS settings and 100% stable when the OS installed, meaning you are not fighting a bad system file problem from a corrupt OS install, if nothing comes of what I posted in suggestions above I think you should go back to x32 and run the clock you want stable and use the 3GB USERVA edit in the boot file to avoid any OOM errors.

At this point the numbers have all been right and they were stable in x32, and, x64 is more stable than x32! So what ever this is it appears to be either OS corruption related or something in that system simply does not play well with x64


If you wish to stay on x64 and run 1600 7-7-7-20 1T PL6-if possible (DRAM STATIC READ DISABLED) thats up to you.

Going over this in my head I just do not see anything in a wild-card here and if none of the items I suggested above has any effect then I really do think this is either a bad OS install or something in the hardware not playing nice on 64

NickN wrote on Jan 12^th, 2009 at 12:01am:


Will do.

vgbaron wrote on Jan 12^th, 2009 at 12:50pm:


   

I'll check the seating out tonight b4 I swap it. Thinking about the monitor click - altho they are not connected, the HAnns-G monitor has built in speakers. Wonder if that click might be an electrical pop from the speakers.

Altho it sounds exactly like the click from the old CRT's when they changed modes.

Further testing tonight.

1) check card seated in slot correctly
2) disable memory map
3) switch monitors
4) switch vid cards
5) have a drink or three

NickN wrote on Jan 12^th, 2009 at 1:48pm:

ok - raised Vtt to 1.31 and dropped GTL to -45. FSB at 430. Seems stable in OCCT *AND* 3DMark is 19900.

Question: Memory is at 1721 8-8-8-24-2T. Can I tighten this up or would it not be worth it. 1600 7-7-7-20 worked fine as did 1800 8-8-8-24

NickN wrote on Jan 12^th, 2009 at 11:11pm:


OOPs - not what I meant. At the old settings 1800/1600 were fine.

Should have waited on those drinks..

Yes - 1721 8-8-8 was not stable in 3DMark until VTT1.31 GTL -45 NB1.55.

LOL - to me -45 is lower than -40.

No I did not try 1800 at -45 1.35 - I will tho

Currently -

430x9  1721 8-8-8-24 1T PL6  Vtt 1.31  GTL -45 NBGtl 60  Vnb 1.44

1 hour stable OCCT and 3DMark 19900

Everest is almost 11000 read with 48.7 latency

n uther question:

at 1721:

8-8-8-24-1T PL6 or 7-7-7-20-2T pl7 - is one significantly better?

Will try 1800 with higher Vtt and see if 3dM flies.

Edit:  no joy at 1800 Vtt 1.35 GTL -45 NBGTL 60

If that fails.. then set back to 8-8-8 and start nudging FSB up 2FSB at a time

When you find the unstable spot (in example 238) back it down 4Mhz and that should be the high note.. you can try adding in a touch of VTT or NB and see if the instabilty firms up but if it doesnt, just back it down and lock it.. remember it may still show unstable in FSX, you know how that works so if it is, back it down till its stable in FSX too.

The only thing you can do now is to raise FSB

NickN wrote on Jan 13^th, 2009 at 12:23am:



Geting close -

430 1721 8-8-8-24-1t PL6 Vtt1.35 -45 +60 nb1.55

OCCT stable - 3DMark stable

Flew FSX for about 35 minutes doing lots of aerobatics in LA basin - lots of terrain rolling by etc - just as I was landing - graphics went garbage - coarse and green almost stick figures - hit alt-enter to get into fullscreen and graphics cleared up -landed with no issue or other graphic anomaly.

How high can I go on Vtt? NB can go to 1.6?

If that doesn't clear it up - I'll drop fsb a tad.

Goping to knock off for the night - Glenfiddich & Amaretto on the rocks catching up.

I bought the BFG GTX280 OC2 which is factory overclocked - I've done nothin to it.

System did not lock up, sound was fine, just graphics turned coarse green and blue.

Bufferpools is set at 70000000 ( 1M card )

Ive called it a night - hope your flu is easing up - I've still got a lingering cough but feel ok.

will do when I get home tonight.

Glad you're on the mend!

LOL!


I have never seen any error like that before


When you back the memory down do you get any of this?

Drop the FSB down to 420 and see if this still shows up

Strange

I have seen a lot of errors in FSX but those are a first for me


Looking something up here....

Set VTT back down to 1.3 and CPU GTL's back up to -40 too

If you see the error again @ 1600 could try swapping the Corsairs with the mushies since the settings are exactly the same @ 1600 and check it again

but if that error comes back I think there is something else afoot here

vgbaron wrote on Jan 13^th, 2009 at 11:35pm:



Alright

Something just does not jive

Im going to throw out a wild card

Crank it back up to 430 and 8-8-8-24 1T PL6 (I thinks that is what you had)

VTT 1.10/ Also try 1.20

CPU GTL  +70  (all 4)

DDR3 ChannelA Vref: +25mV
DDR3 ChannelB Vref: +25mV

Make sure AI Clock Twister has been on LIGHTER this entire time too..

I dont know Vic but thats why I shifted numbers back to the old math because this was just off the wall and not making sense

Leave everything else the same and pump it back up to 450

1800 8-8-8-24 1T PL7 (PL6 wont be possible at 1800)

on VTT you can go as high as 1.30 but start at 1.1, 1.2, 1.28

+70 represents .63x in any other BIOS with VTT 1.10 (1.10v x.63= .693) if the BIOS math is using the standard method of calc and then does the rest behind the scenes.  Engineers know that is not really the value on the surface but its easier to deal with for the typical user and just let the BIOS do the rest. That is how they originally made the RE to work.

If you set it back up to 1800 and find the stable VTT, try dropping the 2 DRAM lines back to AUTO and see if it remains stable.

If not enable them again at the same values

NB GTL by default is +62 and we skewed that back to +60 which should remain the same for what you are doing now.

Let me know if this corrects the stability issue

I dont see that change affecting SATA ability but god knows WTF they are doing with these BIOS's now  LOL


Grrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrr

I dont know if no news is good news or not... but I will check back in a bit

bummer


OK

Then at this point and assuming 1800 was completely stable in the 32bit OS I must assume there is something in that system in combination with 64bit which simply does not play right

Right now I am just using EVGA's Precision Tool on the i7 system with the 280...but thats not an auto fan control, its manual.

I have not had any time to play! I cant build GEX on the i7 system or the perf would be out of line with typical systems so I have not spent a lot of time with it setting up all my auto controls. I will probably set up RivaTuner on it when I get time but I have not checked to see if they updated it to control the 280 yet


I guess you are going to have to go back down to 430 8-8-8 vic and set the numbers back up as they were stable on that clock. The benchmark numbers were not that far off the 1800 values so I dont see you losing anything on that clock, thats for sure but I did want to know if it would stabilize @ 1800.

I could probably sit here and throw skew values out at you for 2 days and hope for a winner but I'll tell you truth the amount of going back and forth and you testing for another sliver of perf, I just dont see the value other than the loss of 180MHz CPU clock, which is really nothing in the big picture

I agree. 430 8-8-8 seemed just fine and the bench differences were just that - bench differences - nothing that I could translate to FSX.

Just tried Riv2.2 and it doesn't handle the 280n right for fan control. I'll revisit it JIC it was me misunderstanding something.

I want to get some flying done so I'll just leave it at 430 and if something else crops up that's an issue I'll drop it down a bit.

Not going to buy any more hardware until the I7's come down. Of course, by then I may have to deal with Windoze 7.

I will try VERY hard not to come up with any more brainteasers for a while - I'll just stick to simple things like nHancer settings, etc..

As always, I appreciate your help and certainly enjoy the banter - look forward to doing it over a cool one some day.

Also, hope readers of these threads learn a bit about how things can go.

Now to go get Acronis True Image - need a decent b/u program for XP64.

Thanx again -

OK well then THATS OUT too

Glad you did that rail removal.. they were known for being pooly fitted

Riva is not working the 280 fan yet?  crap.. well, I dont have to sit here and listen to it all day so thank god for that. I hope they get the support for that nailed down soon.

you could go back to x32   

I didnt think so....   

little secret

x64 is smoother than x32

and XP x64 is smoother than Vista or Windows7 x64

so you doin fine and on the right OS and hardware, for now