test reply. since i just lost my previous post.
Printable View
test reply. since i just lost my previous post.
OK. typing again.
I have a XFX 790i Ultra board rev. 1.1 with Bios P03. but running it with a HIS Radeon 3870X2, , OCZ 4GB DDR3 PC3-10666C6 1333MHz ReaperX (2x2GB) Dual Channel DDR3 (OCZ3RPX1333EB4GK), E8500 and Coolermaster 850W PSU.
The problems I have on a stock setting system except the memory voltage and unlinked mem settings are as follows:
Raid Problems:
Occassionally at boot the sata controller can't find the stripe
the Nvidia 9.64 sata raid service claims there are errors. However using the seagate dos check tool for 2 hrs per drive finds no problem
Black screen / reboot
Very occassionally in BIOS, but frequent when Vista X64 has booted I get a the strangest of the times the effect that the Graphics card fan speeds up, and gives a black screen / drops the signal to the monitor. Initially I had the problem that although the PC remained on, the rest of the system was dead. E.g. hitting the numlock did not do anything. However, y'day I noticed when i got the problem again when loading Hellgate that actually the keyboard still reacted, the sound card produced sound and i saw the hdd still being active.
Please note that i have already rma'ed the board once here in the uk with the above problem. the retailer is now suggesting to rma the board and the cpu.. but based on all what i am reading here, not sure if that solves the problem.
my question to you guys is as follows:
do the raid problems described by me others are experiencing as well? based on the 17 pages it seems to be so. although after 15-30 black screens i am still able to restart vista, but never tried the scan / ... thing will do that later today.
should I move away from the graphics card / motherboard and is that the rootcause of all evil or am i overlooking something else. e.g certain bios settings / voltages etc?
if I move away I can either decide to move to 9800X2 or for example move to a X48 Asus P5E3 .... would want to stay on a DD3 board..
Any suggestions? any additional info needed?
try everything at default speeds to make sure its not a configuration problem. if it works there, you gotta find out what to tweak to get it stable at higher speeds. if you have the same problems there, then swap hardware if possible to troubleshoot what causes the problems... but it sounds like its the mainboard to me...
either instability/configuration problems, or its the mobo...
Thanks,
at present I am not even trying to get at higher speeds. or are you saying that since the 790i tends to have the default bios to DDR 3 at 1066mhz rather than factory default at 1333, I should try it at 1066?
I just checked the XFX force website, but they don't have the P04 bios yet. Ati has released a new catalyst driver y'day so will use that one as well.
But since the sun is now shining in this country called UK, yes I know hard to believe. Gonna take a stroll outside first ;)
you can use evga P04 bios on xfx mobo.
When I tried to flash my XFX board with Evga's P04 BIOS it wouldn't because the part numbers didn't match .. This was done through the AWD Flash utility with the file burnt to a CD as an ISO .. Is there a way around this? Not through windows tho (i've lost boards before that way)
Make a bootable CD (or a floppy if you use FDD), write on it your BIOS file and awdflash files, type awdflash /f NF79_P04.bin manually, it will be done.Quote:
Originally Posted by Anubis_386
No raid problems with any BIOS I have used, I am running 2xRaptor 150Gb under Vista 64. Your problem looks like a hardware problem / bad connection.Quote:
Originally Posted by Desarts
I had 780i and the same raid partitions work fine with 790i.
Take a look at your SATA connectors/cables, are they all fine?
Try to run memtest86 for several hours - will let you know if CPU/memory work fineQuote:
should I move away from the graphics card / motherboard and is that the rootcause of all evil or am i overlooking something else. e.g certain bios settings / voltages etc?
Try to run HDD tests. Good utilities can be found on Ultimate Boot CD (UBCD) and on Hiren's Boot CD.
Your need to troubleshoot your system, start with the default settings (load defaults in BIOS) and make all tests.Quote:
if I move away I can either decide to move to 9800X2 or for example move to a X48 Asus P5E3 .... would want to stay on a DD3 board..
Any suggestions? any additional info needed?
Also flash 04 BIOS, it looks quite good comparing to 03.
Quote:
Originally Posted by saaya
My fault, I really meant C0 instead of C1 :p:
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.
That scared me for a second but since it hasnt been mentioned again, I really dont know if there is going to be another stepping or not.
Can't wait to see what happens with the SII E.Quote:
Originally Posted by BenchZowner
I've finished the "stock" measurements for the review, at FSB 335MHz and the mem linked & sync mode at tight timings ( 5-5-4-15-1T-1-1-1-36-1 ) with no obvious signs of data corruption.
Now I'm tweaking stuff for the overclocked measurements, but to go fair for a clock per clock comparison with P35 I'll use 400MHz FSB & Linked/Sync mem again.
I'll have the measurements done in 6 hours or so, after that I'll start torturing the S2E with extreme clocking, etc to see what can be done about the corruption issue.
I'll try to find the source of the problem ( what's causing it ) and see what I can map/do about it.
Im really glad to hear that BenchZ.
Show us some 500mhz FSB benchies with BIOS 0601 ;) (with no corruption of course)
I am afraid corruption is still there.
I made a fresh install of XP 32bit since I flashed 0601.
I run at fsb 1800 linked and synched. I had memtest for over 2 hours and 8+ hours of CPU Stress MT, everything stable.
1 hour ago, out of nowhere, I got a BSOD "bad pool caller".
I rebooted and got the message that the registry was restored from an older version. Since then, I was unable to install games because windows installer service could not be found.
I had to restore XP to some point prior to my BSOD to get everything right again :-(
Now I installed vista 64 to investigate further about corruption issues.
I did the install with my oc at fsb 450 to increase the chance of corruption to occur.
Right after the install was over, I tried "sfc /scannow". No corruption.
The sfc program is so much better than in XP since it actually tells you if corrupted system files were found or not.
NOTE: "sfc /scannow" doesn't say anything about corruption in XP apart from asking you about the CD. And that doesn't mean that you have corruption IMO.
I tried to run sfc on my old rig (also on XP SP2), and it also asked straight for the CD when I know I have no corruption there.
Also I tried to run sfc with XP SP2 CD in the drive so that it would supposedly repair everything. Right after it was done, I ran sfc again (I removed the CD just before that), and before even scanning the first file it asked again for the CD!
What the hell?! I don't think sfc is the way to assess corruption in XP 32bit.
OS is now corrupted. Was running so well there for a while, but Vista got borked when i made the mistake of suspending instead of shutting down, on resume the system hung, bypassing resume lost all my user acct settings/data. Ran SFC and got lots of corrupted files.
any other way to run a system check for corruption instead of sfc? a program or something?
This is where im at so far, no corruptions that i know of, rock solid.
http://img516.imageshack.us/img516/5517/setupjc9.th.jpg
It looks like a hole at 450MHz bus (1800), I had some instability issues and moved a bit up to 452MHz (1810MHz), now everything works fine and stable, no BSOD, nothing.Quote:
Originally Posted by MateoTTR
Well, that's the thing, I am very worried about this board now. I had zero issues for weeks, rock solid. And now, one day, all of a sudden, I got corruption. This is not a good sign, and I really hope the same doesn't happen to you guys who have been running solid so far, as I was.
Happy to report that I defragged 1.4TB of mostly 8GB+ size files yesterday across 3 drives simotaneously without issue :D
G0 Q6600 @ 3555MHz
444x8 / 1777FSB Linked n Synced
7-7-7-21-1T
The fact that many people are having issues with this chipset is very disturbing and indeed not the first time this has happened with an NVidia chipset.
It'll be a frosty day in you-know-where before I use one of these boards in my system, as my data's integrity is the most important thing, indeed the only thing that truly matters to me. With the amount of data in my computer, data corruption would be an absolute catastrophe. If NVidia can't produce a reliable chipset, they should avoid the server or workstation environment entirely...such issues just aren't tolerated.
Hello,
following the suggestion to do component testing and upgrading to P04
upgrade my xfx 790i to PO4
All is default except:
put floppy present to no
enabled raid config
disabled HD Audio
Put Vmem voltage to 1.85V
Put FSB to Mem as linked and synch to get my memory to the right 1333 stockspeed.
Timings are set to Optimal by Bios itself, resulting values are putting it at 8 7 7 25 1T (offically the memory OCZ repear is 6-5-5 18 1 T)
yesterday and ran memtest for 2-4hrs. This passed without any errors.
Using now the UBCD v4.1.1 and ran the following tests so far:
2+hrs of memtest - pass
StressCPU :with default settings (12.25MPFLOPS) results in a black screen after 375000 iterations, the system gives then actually a post boot code of 98 --> checked the manual, but it does not have any postcode like that.
Key question here is: is the number of iterations of 375000 a normal result, or far too low? If the test fails at the end of the day, does it not normally result in a black screen or? Any one knows this post boot error code?
Update: Actually whilst the StressCPU runs, the mainboard is actually alreaydy giving the alternate flashing of 96-98 as a post boot code... So the result actually is that the system at X iterations just gives a black screen, but the monitor signal is not dropped as is my original overall problem.
i have just had my first issue with my xfx mobo. i was browsing internet when it suddenly shut down. i couldnt power it up again. tried clr cms, battery remove but nothing helped. after a while, trying power it up, i could power it up. the diag said FF constantly. i didnt give up when finally it posted. i didnt do anything special for it, but now it works. its a x-file for me :(
my config:
XFX 790i Ultra P04
QX9770 @ 4.0G
2GB G.Skill DDR3 1600 @ 1600
Asus 9800GX2
Ultra x3 1kW
150GB WD Raptor
320GB Seagate 7200.11
settings:
fsb = 1600
mem = 1600 sync mode 7-7-7-18-1t 1.80V
cpuv = 1.42v in bios
fsb = 1.30v
spp = 1.40v
mcp = 1.55v
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.
http://img530.imageshack.us/img530/5...0268tk0.th.jpg
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 :shrug:.
Describe the freeze please?Quote:
Originally Posted by tekjunkie
System was hanging. No response to mouse or keyboard.Quote:
Originally Posted by 5010
Sounds like the same freeze encountered by oh so many others including myself until I balanced my FSB & GTLREF properly.
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 GTLsQuote:
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"
Dont say that.... I have a 2Gb HZ kit on my hands already! :confused:Quote:
Originally Posted by tekjunkie
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?Quote:
Originally Posted by OBR
Upgrade to 0601...
:eek: 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 :DQuote:
Originally Posted by Desarts
nonono, they made a sarcastic comment in regards to c0 ES :DQuote:
Originally Posted by slim142
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 :D
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.
To test 3D stability, I run 3Dmark 06 in a loop for 6+ hours.
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.Quote:
Originally Posted by slim142
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!
Quote:
Originally Posted by eternal_fantasy
:slap: :stick: :slapass:
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
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.
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! :mad:Quote:
Originally Posted by saaya
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.Quote:
Originally Posted by tekjunkie
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.:ROTF:
I got this problem too. Any help?Quote:
Originally Posted by Lu(ky
Have you guys tried 0601? (From pmp)Quote:
Originally Posted by Broly
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... :mad:Quote:
Originally Posted by Broly
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 :)
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.Quote:
Originally Posted by Xello
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.
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.
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.
@gcardinal
i sell my Asus Striker II Extreme and buy an Asus P5E3 Premium. the best decision that i could make. never NVIDIA again.
HA! even you sold your 790 board now? :lol:
you were one of the few people who seemed to be really happy with their 790 boards :D what happened? corruption?
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.
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:mad: And the next moment, my OS corrupted:shakes:
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 :)Quote:
Originally Posted by webwilli
Well, I guess you can use CrossFire :) , So whats the amazing OC you achieved now?
:ROTF: Good Luck with that :rofl:Quote:
Originally Posted by webwilli
I have 0 issues & love my Striker II...:up: :up: :up:
I have completely opposite feelings ... 460 MHz FSB on a quad core, 4.14 GHz on air cooling, 22691 3dmark06 with one 9800GX2 and stock cooling, memory at 7-6-5-18T-1T@1840MHz linked sync ...Quote:
Originally Posted by webwilli
I am waiting for the WC set ... all voltages are in the "safe harbor" ... 1.35V VTT, 1.39V CPU, 2.0V memory, temps could be better, of course, and will be better with WC :yepp:
People who have no issues seem to be running 2gb.
Yes, but that can be said for overclocking on any mobo.Quote:
Originally Posted by Broly
Thats true, but most people running 4Gbs here are running 2x2Gb so I dont think there is an increase of stressing the NB
Apparently the 2x2's are massively OC'd
My patriots are JEDEC 1066 I think, and they're rated for 7-7-7-20 at 1600 MHZ which is a little too good to be true. I think I am going to sell these and buy the new Cellshock or dominators.
Can anybody verify that this corruption issues has something to do with SATA HDD instead of IDE HDD? Or the chances of corruption is higher on SATA HDD as compared to IDE HDD??
Grateful for ES, where we can warn folks when things like this go on.
:(Quote:
Originally Posted by webwilli
of course it makes a difference!Quote:
Originally Posted by slim142
whether the chipset has to drive 2 single sided or 2 double sided sticks makes a huge difference. its double the chips the memory controller has to drive... and the board has to power those chips too and keep vdimm and all the other voltages stable... the more chips, the more load on the vdimm circuicy...
i dont think the amount of memory makes THAT much of a difference when it comes to instability though... sure, more mem is more stress so higher chance the nb craps out... but it rather sounds to me like some boards clock great and are stable at high speeds, while others are not. there seems to be a big difference from one board to the other. which is exactly what happened with 680 btw.
from everything people who had a lot of boards told me there was a rough 1:2:2 ratio of 680 boards. 1 great, 2ok, 2 real bad.
it also happened when people used raid controllers...Quote:
Originally Posted by memory
but iirc yes, people seemed to have less corruption problems with IDE hdds...
[QUOTE=saaya;2956012]
of course it makes a difference!
whether the chipset has to drive 2 single sided or 2 double sided sticks makes a huge difference. its double the chips the memory controller has to drive... and the board has to power those chips too and keep vdimm and all the other voltages stable... the more chips, the more load on the vdimm circuicy...
i dont think the amount of memory makes THAT much of a difference when it comes to instability though... sure, more mem is more stress so higher chance the nb craps out... but it rather sounds to me like some boards clock great and are stable at high speeds, while others are not.
[QUOTE]
Sorry for not clarifying my statement. I meant what is in bold :up:
LOLz!Quote:
Originally Posted by Broly
Sorry but all DDR3 performance kits are just MASSIVELY OC'd Micron D9's :p:
My Patriots will do DDR3-2000 8-8-8-24 on just 2.0V just the same as any other leading kit only they cost less.
Around the EVGA forums Patriot has the fewest issues on the 790iU, Actually "NONE". Been reading all leading 790iU threads around the interwebs since launch and yours is the first Patriot issue I've read and I wouldn't count it as an issue since you seem to not understand DDR3 much in the first place.
Gskill has the most issues and OCZ comes in second.
EDIT: I just realized thats you Broly... ROFLOL... I still remember your tantrum on the EVGA forums followed by all that BS you tried to feed everyone about your tech knowledge....... should have known it'd be you making a comment like that :P: - No offense or anything..
Saaya
add strikerII nse ?? no corruption yet. :)
longsiew!! a lot of times due to GTLRef / VTT Ratio... try 0.584x for ref 0 and 1. i saw a ref mobo's VTT move and those spikes always gave me corruption. if u can monitor vtt throughout a period of time to see how stable it is...Quote:
Originally Posted by memory
I second this theory.. Ive had one case of corruption that was specific to my CRYSIS install a few days after I got the board I know the issue which triggered my corruption was caused by bad VTT/GTL. Since my own experience I've noted several times seeing posts around the web(here aswell) that people could pinpoint the exact time the corruption occured and those times were always after a "freeze-up" exactly the same way 790i's freeze when their VTT & GTLVREFs are bad.Quote:
Originally Posted by LardArse
i know, just do an experiment with variable resistors to change voltage on the fly you'll see straightaway.Quote:
Originally Posted by 5010
Is this for quad. Going by your reference to both 0 and 1, I assume that it's for quads. Also, I though it was .66 times the VTT is the reference. So if that's the case, do I have to go down on the GTL(-xxx mV)??Quote:
Originally Posted by LardArse
Heh, nice to have someone finally confirming my suspision. Far to many folks are having issues, 99% have more money than brains and just can't handle VTT/GTL theory, yet its so necessary w/790iU even at stock speeds the auto VTT & GTLVREFs are totally borked and just cant go unadjusted. Hence we have 1000's or cry-babies who cant or wont help themselves to solve the problems their having with the board posting around complaining everywhere and spreading false fixes w/ driver reinstalls etc etc.. A real mess!Quote:
I know, just do an experiment with variable resistors to change voltage on the fly you'll see straightaway.
.67x is 00mv(standard) in the 790iU BiosQuote:
Originally Posted by tekjunkie
65nm Quads like it abit higher; .68x - .70x
Duals + Yorkies like it lower; .66x - .63x or even less (as low as .58x according to LardArse)
Quote:
Originally Posted by 5010
Thanks. I went with what AT had in their S2E and always had it in the postive range for my QX9650.Unless I read it wrong, which is quite possible these days with me staying up all night playing around with my s2e, they had said that the quads require more than the reference,i.e, postive GTLs.It didn't do squat for my stability. As a matter of fact it was detrimental to the stability. I am SPI 32M stable @450FSB with my QX9650. I am gonna run prime now. I will prime with GTL @ auto and @ 0.63 times VTT and post back with results.
I had corruption issues with P03R2 BIOS before with the same 2x1Gb, it's not a point. The voltages idea looks more appropriate ... the auto correlation could be corrected in P04.Quote:
Originally Posted by slim142
All but the new and NYA Corsair 1800 4G kit which uses Samsungs.Quote:
Originally Posted by 5010
ref: http://www.xbitlabs.com/articles/mem...-1800-4gb.html
I would really like someone to get them (b4 me) and play with them on the 790i boards.
I decided to try the 790i after abandoning nvidia chipsets after the 680i problems and yes I did have corruption problems, in fact it led me to mac pro after years of building my own systems. I just got tired of all the incompatabilities and stuff just not working.
I have finally completed Striker II Ex rig (fully watercooled by two Swiftechs), Q9650, 2x Ultra in SLI, 4GB DDR3 1800MHz and Raid0 from Raptors and with BIOS 0601 in last week no problem ... no corruption, no issues ...
Great stable system on 450MHz (RAM 1800 - 20-7-7-7) CPU 4050MHz ...
dont you think this is simply based on coincidence and hence the more widespread a brand is the more people will report issues using this mem?Quote:
Originally Posted by 5010
ocz seems to be very widespread and gskill as well while patriot is used by very few people.
your the only person i know using one... and i bet its not a retail board, or is it? :DQuote:
Originally Posted by Rol-Co
yeah, the 790 problems all seem to have to do a lot with vtt.Quote:
Originally Posted by LardArse
that explains why some boards have this issue and some dont, and why some people report a board works fine with one cpu and doesnt with another. the problem is that 65 and 45 and quads and duals have different preference for vtt, and even the same type of cpu has different versions of cpus, some that like higher and some that like lower vtt...
ideally the board should set a certain vtt automatically depening on whether a cpu is recognized as dual or quad, and whether its 65nm or 45nm.
OBR, why do you keep posting that you dont have any issues?
first you said all was great, again and again, then you said you had instability issues too and didnt post for some time, and now your back posting again and again how everything is rock stable :P
instead of repeating how your system is stable, which doesnt help anybody, could you please post your detailed settings or maybe some tweaks you found to improve stability? that would actually help people... just telling people that have stability problems that you dont have stability problems... isnt really helping anybody :D
I assume that the value you calculate is then applied to both dies on the data bus..so why are there different values in the bios for REF0 and REF1. Using LardArse's .584 at 1.36VTT I get a -117mV adjustment, but this value is not available for REF1 adjustment (-120mV is the closest), although it is available for REF0. Am I missing something, or do I just get as close as possible to a value in between?Quote:
Originally Posted by 5010
no i had problems (after some weeks of using), but now with last bios are problem gone ... between first problems and now was SIIEx in the box ... now after one week of intensive using - NO PROBLEMS at all ...
I´m having a hard time with this board (S2E) :( . Just cant overclock without any issues! :mad: I´m having data corruptions too. I have 1866Mhz Cellshock memory that can only run at 2T at 1600Mhz. If i try to put 1T it´s gonna give BSOD!!! Can´t overclock my QX9650 to a ridiculus 400MHZ FSB. I´m just very disapointed with this board. Can anyone give an idea of the temperatures for the NB showed in BIOS, because mine is at 55º~57º at idle. I dont use WC and i´ve not changed the TIM in the chipset.
Quote:
Originally Posted by saaya
Quote:
Originally Posted by OBR
OBR Instead of saying your stable not stable every other post. How about your BIOS SETTINGS that got your stable to help others that are on this thread? Remember that is why there is a Forum and Threads to HELP people...
Also what DDR3 4GB kit are you using?
Thank You
I share your frustration!!Quote:
Originally Posted by Lu(ky
There were tests done to test the tolerance range for reference voltages on the chipset, the range is much smaller than on an Intel chipset.
As signal strength varies according to temperature of chipset/cpu, VTT/GTLRef values have to accompany the temperatures so that is a problem.
Aight so here are some updates
Got tired of all of this drama and Im giving a try to the Striker II Extreme (Yeah ya all heard me right :P )
Will see if this motherboard can challenge me face to face
.63 times VTT for GTL woked like a charm :up: for my QX9650. Better yet, it enabled me to lower my vcore. woot. My VTT is @ 1.32 for 450x9=4.05GHZ and the GTLs are @ -48mV and -51mV for the data lanes.
This doesn't sound like something we are going to see fixed with just a BIOS update. Not good, not good.... 100% stability is still possible by extensive tweaking but you've gotta be awful lucky to do a build with the right cpu/speed combination to get stability out of the box. :shakes:Quote:
Originally Posted by LardArse
I really love the board but I'm not your avg Joe, avg Joe's want plug and play and have no idea what GTL is, let alone how to properly adjust it.. Gonna be a fun time to work in the RMA dept @ EVGA, XFX & ASUS etc..
the temperature of the chipset is a mayor factor in getting signal stability?Quote:
Originally Posted by LardArse
and the cooler the better? does the signal window widen up the cooler the chipset gets? 30 is xxxx, 40 is xxx and 50 is x?
or does it just change depending on the temperature? and the singal window moves with differnt chipset temperatures, at 30 its at xx, and 40 its at xy and at 45 its at yz
tweaking this must be very frustrating... since there is still no reliable way to detect 100% stability. it seems that just putting load on the chipset doesnt trigger the issues, varying states of idle and load seem to trigger the problems... and simulating that to test stability is a pain in the arse....Quote:
Originally Posted by 5010
so at any time you might think NOW you got it stable, but then it fails again after 2 days... what we need now is a propper detection scheme or benchmark or tool.
the scf scannow doesnt seem reliable as we get many false positives
several people reported that streaming a lot of data triggers the problem, has anybody used pcmark or some other overall system benchmark that tests each part of the system after the other? this should result in repeated idle and load situations on the chipset to some degree and might trigger the instability/corruption issues quickly.
once we know how to trigger and detect it quickly, we can find out how to improve the situation and get rid of the problem id say.
maybe we can come up with a guide on how to tweak vtt on 790, but seeing as its a different thing for 45 and 65nm cpus and quads and dualcores... and each cpu and board behaves different... this guide would have to be quite complex.
i've done it with both IDE and SATAQuote:
Originally Posted by memory
i don't think that's got anything to do with it
maybe shamino is onto something but until tested it's just another theory. It's great Peter is taking the time to help people out here....whoever is feeling confident finding those voltage points try to test these things out and let us know how you go. I've started a general voltmod discussion thread here so this could be a good start when it comes to finding the mods required i guess :)
http://www.xtremesystems.org/forums/...d.php?t=181677
if anything i'd say 5010 seems to be the only person so far that has touched onto something we can go on in this thread >>>> the frustrating part is that there are PAGES of :banana::banana::banana::banana: in this thread now and none of the people voicing frustrating have bothered to read what the guy posted or bothered to test the settings out
There is a reason why overclocking is a skill....you have to spend lots of constructive time honing and getting to the bottom of a problem at hand.....
i've had many PMs so far about this kind of stuff and i really dont think anyone can really give ppl magic settings in bios and automatically make everything better.
Everyone read back what 5010 said about GTL Ref and notes be posted about his testing and then test out GTL on your PC >>> and i dont mean go copy 5010's settings but actually TEST OUT ranges your hardware configuration works best with
saaya you need to take charge of your thread mate since you are trying to make a contribution. I think it's great you've gotten up and started it all but you need to give people direction in your thread. Right now it looks more like an OC discussion that you would have in a pub with your mates after about 6th or 7th round of drinks :p:
Personally this corruption issue i can avoid when i bench as i know that all i have to avoid is not push RAM too much out of it's comfort zone so i just stick say 2GHz 8-7-6-20 1T with 1.95v and i'm fine but i do want to see people sort this out of course as this is such a great board aside from this issue in my opinion :shrug: /RANT
not really u just need different levels of volts with different temps, even cpu tempQuote:
Originally Posted by saaya
whaaaaaa?? i know what you mean but hows it my fault or responsability to moderate this thread or tell them what to post and how to post here?Quote:
Originally Posted by dinos22
if id still be a mod, sure, id mod this thread and remove OT posts or edit them...
but just cause i started this thread doesnt mean im responsible for what people post here :stick: :P
I ask people to post more details if they dont post any details, i try to help everybody by suggesting settings to try, what else can i do?
am i supposed to btch at people cause they dont listen to what i tell them? what would ive me the right to do that? just cause i started this thread and its MY thread? i dont really consider this my thread as in my propperty or responsability...
and no offense, but YOU have posted a LOT of OT stuff here... ;) :hehe: so dont you point at ME for this thread not staying on track all the time :lol: :D
I know your exhausted cause you just arrived back home from a week long trip, but that doesnt mean you need to poke at me cause you dont like a thread i started :D
hmmmm but a lot? or slightly?Quote:
Originally Posted by LardArse
So a setting that worked on a cool day will not work on a hot summer day?
or a setting that worked under ln2 will not work on air?
the latter is what i expected :D
if its really that bad and 10 degree temp difference change things then... this is quite frustrating...
lol in know man
you and me are OT kings hahah
but seriously though
ask for mods to clean the thread up and use the first post to keep the thread up to date with possible workarounds and some testing updates if possible that you could copy and paste from useful posts here :shrug: :)
meh, why me? :P
cause i started this thread?
i just bought assesins creed and imma spend my weekend with that, and not trying to get peoples penis fights in this thread removed or read through 40 pages to collect infos :P plus i dont think any of the mods is willing to clean this thread up...
maybe next week... if you want this so bad why are you pushing me to do it?
go ahead! send me some roundup text to add to the first post and ill add it :D
Oh Hmmmm.. eh
If I get any problems, you guys will have to explain me a little bit this .63 times VTT for GTL, -48mV and -51mV for the data lanes and all that mV stuff because I never touched that and since it solved problems for 5010 and tekjunkie, I believe that might be the Key we been all looking for.
Well my Asus Striker II Extreme finally gave up on me today... I only had 3 days left on this board through the Egg or I would have had to send it for a RMA through Asus :clap:. The girl asked me if I would like a REFUND or a RETURN. I said how about a FULL REFUND no restocking fee.:woot: She said no problem send it back to us and we will give you a full credit.. :) Not to bag on this S2E but I am no that impressed coming from a eVGA 680i other then the DDR3 and the boards layout is great. I hope things improve for you all I will be moving over to a x38/x48 board to try out. See attached pictures of my problem I got about 2 hours ago when I turned on my board. And yes I changed out 3 GPU's, 2 CPU's, 2 PSU, and original 2 x Raptors 150 raid -0- with one Raptor in Windows same thing... I plugged in my old setup and it is not my monitor.. So it is the mobo for sure.. Good Luck all if things turn around here I will be back for sure... :up:
http://i274.photobucket.com/albums/j...n/IMG_0922.jpg
http://i274.photobucket.com/albums/j...n/IMG_0919.jpg
Im glad to see you are so calm about it anyway :) creds to that, i would have gone crazy :)
could've been all my music on that raid, or something.. that i collected for years, sure its not smart to keep that on a system disk, i sure dont, but some might :)
i really got interested in this 790i in the start when peps showed high clocks, high fsb etc :) and COULD GO SLI , but sooon this corruption came, and wooo hooo , noo im getting this X48 :)
Ah well good luck, hope you find yourself back to Intel :P
So Newegg will give a full refund?? Man, I've RMA'ed mine 3 times. I'm running Prime while I'm typing this. If I had know I would have asked for refund from day 1!!! This time I left outside the case, just waiting for the headaches again. It didn't crash for the last 20min. Oc'ed my E8400 @ 4.05Ghz and ram OCZ EB1600 to 1800 8-8-8-24 2T. 1T just freezes Memtest. Man, right now I'm kinda wishing for corruption or BSOD. It would be awesome to return it and finally move on. But if this thing give me errors after 5+ hrs of Prime, I swear I won't have anything to return. I'll destroy it for good.Quote:
Originally Posted by Lu(ky
Erm, it would be nice if the people who know what theyre talking about could post it anyway. Ive tried so many VVT voltages and GRL ref voltages over the last week and a half, and nothing at all is stable. Im only running 1066Mhz DDR3 RAM (yeah thats right, the motherboard stole all my RAM money), but i am using 4x1gb sticks. Using 2gb and only one 9800GTX does delay the inevitable freeze but its just a pain im getting more annoyed about day by day.Quote:
Originally Posted by slim142
Im just completely stuck, can anyone give me some starting points?
Oh, no corruption of stock speeds either, which i suppose is something. But i cant play games, so theres no point of really OCing at the moment anyway.
My system's been running just great, did a bit of a bench with my q6600 last night.
http://img232.imageshack.us/img232/8...k06rlm4di8.jpg
However, look's like I've just got my first taste of the video corruption issue.
If I change my ram to 1T 452 x 8 or 452 x 9 synced 1810mhz 8-8-8-1T, then 3dmark 06 fills the screen with garbage and eventually blue screens with the familiar nvspdll error.
At these settings I pass memtest, and albeit a reasonably short run, prime blend doesn't throw up any errors.
Up to now I haven't had to change my GTL refs, auto has carried me through no worries. Settings in my sig are nice and stable.
Just wonder, those who have been pushing this board, Dino and the like, have you come across this issue and did you find particular tweaks helped?
Cheers
RLM
I agree. This is the first time I see that a motherboard "might need" (since RFL didnt need to change this) to get the GTLRef voltages changed and those mV stuff so Im kinda new on all this new settings that I never changed in my life. I just hope this motherboard doesnt get me to suicide.Quote:
Originally Posted by Coupa
Once I get my motherboard I will spend around a week (entire week) to see what does it like, what it doesnt like, how it performs and everything. If I get this nightmares to the point where I will get an email machine with SLI, I will send it back even when this is the most amazingly and pretties board I ever seen (physically)Quote:
Originally Posted by golfjam
RLM, do mentioned you had blue screens if you changed to 1T, dont you think that already messed your OS installation and might be the cause of the BSoD from the display driver? or probably corruption somewhere else is leading to the bsod you are experiencing.Quote:
Originally Posted by RLM
Im glad to hear you are somewhat progressing through this and not having to mess with the GTLRefs. However, something tells me this might be a a good key for system estability whether we like it or not.
The reason, I think, that you didn't have to mess around with your GTL, could be because of the fact that you are running a 65nm chip. Wolfies and Yorkies, as far as I have, read are more senstive to VTT fluctuations and GTL becomes crucial for 45nm chips. Also, what frequency are you running your PCIE lanes. I refuse to OC my PCIE lanes, because of the corruption problems that we used to have nforce4(No I didn't have a nf4 board, just read about it). I always try my best to keep it @ 100 and I don't bench(eventhough I am ruuning my board outside a case, but I am looking at the new LLs). Nice OC BTW. Good luck.Quote:
Originally Posted by RLM
I cannot believe!!! It passed Prime and OCCT like nothing bad ever happened!!! And the only thing I changed was CPU voltage @ 1.3v (1.276v after vdroop) and ram (manual timings 8-8-8-24 2t @ 1.9v). Everything else is on AUTO. I thought I had to battle to find the right VTT and GTL settings but IT's WORKING!!!! And no corruptions (so far...). Please Lord, don't let it randomly lock up! LOL!!
I have tried possibly all combinations of vtt, gtlref, nb, sb, vcore, vdimm to try and get my rig stable at fsb 1800 linked and synched.
At several occasions I could prime and memtest for 10+ hours, so I started to believe that I was rock stable.
However I was unsure whether I was subject to data corruption or not.
After a week of running "sfc /scannow" on a daily basis, my system files were all of a sudden found to be corrupt :-(
So I reinstalled vista x64 and decided to install a few games to test sli performance.
Wanted to try Assassin's creed and I was surprised to get crc errors during the installation process! Most often I would hit "retry" and the same file would be extracted again but with no crc error.
Sometimes I could install completely the game without any crc error, then would uninstall/reinstall and get errors again, somewhat randomly.
I went to stock settings and fsb 1333, never had a CRC.
Back to fsb 1800, tried many combinations of vtt, gtlref to realize that the occurence of the crc errors was very sensitive to these settings. At best I get the crc error only once every 5 installation of the game, but can't really get rid of it.
I made an iso of the game, to run the install test faster with daemon tools.
Well all this to say that corruption pb exists, it is very difficult to reproduce because you need a combined stress of CPU, RAM and HDD, but it seems that extracting data from an archive and testing crc is a much better/faster way than just waiting for your os to get corrupted.
Hey MateoTTR
Try tekjunkie settings, he also has a QX9650
QDR/RAM=1780 Unlinked
PCIE1/PCIE2 = 100
SPP<-> MCP= Auto
vCORE=1.375(1.344 in windows per cpuz)
Loadline =Enabled
vTT = 1.28
PLL=Auto
vDIMM=1.9 (Manually put the stock voltage for your ram)
vSPP =1.46
vMCP = 1.55
CPU GTL_REF0 = -29mV
CPU GTL_REF1=-32mV
You can also trying not-even FSB speeds like 1805 or 1810. See if that works