**Official DFI LanParty UT 790FX-M2R Review/Overclock/Guide Thread**

[Rammsteiner]

Hmmm,

Ok guys, Ive some weird things going on.

I had some crashes while gaming. At most once a day though. I suspected my sound or videocards.

But as the ISP messed up once again yesterday I decided to have a look at the crashes and it's most likely from my RAM. So I played and played around. Drive Strength settings had the most effect. Timings (subtimings mainly) didnt hardly do anything, Voltage increase neither.

Prime95 ran at settings x 17~20 minutes before error, at settings y only 1 minute before error and settings z like BSOD. But there was nothing to make up out of it. Those settings were mainly between Drive Strength @ auto, something worked before and a mix from my succesfull DDR1140 settings.

So I left the Drive Strengths as I had them at the very first time, 1.0, 1.0, 1.0, 1.5, 1.5, 1.5, decreased the Vdimm by 0.05V and increased Vcore to 1.5375V. All subtimings on auto. Ran SuperPI 32M without problems, changed only MAL from 8 to 4 and tRC from 24 to 16 after that and ran it again without problems...

I didnt try Prime95 yet as I cant be arsed after 2 hours of BSOD's testing if it's fixed or not. But the fact it ran SuperPI 32M stable and even with decreased Vdimm... Not sure but Im pretty sure my RAM wouldn't run with lower Voltage before. That was quite suprising for me.

Also as it's 32M stable I think it's pretty close from being stable. Though stupid enough I didnt test with 32M when the BSOD's occured in Prime95.

Any idea though why at once it wasnt stable anymore? It's quite strange IMO. I dont think it's due to D9's, otherwise the decreased Voltage wouldnt have done anything I guess.

[Oliverda]

So, the 3/12/2008 BETA is the best BIOS for Phenom so far.



VALIDATION

Unfortunately it's unstable but it can boot up at least! I'm working on stable 3GHz right now.

I've found a strange bug in the AOD (both 0.14 and 0.17). It sets back the VCORE to the default 1.25V when I start it.

[Rammsteiner]

Whoa, very nice OC!

I cant wait for B3 to get released and if it does any better. After all I bought a X2 just to wait for a more mature Phenom . Any benchies to show off?

Anyway, a little update, I dropped the HTT by 1Mhz, increased MEMCLK Drive Strength to 1.50 and it's right now priming for a solid 2 hours. Previous attempt was only 54 minutes, then MEMCLK was only at 1.25... So maybe I can push the HTT later on back to 251 if this is stable. Though Ive very little doubt it's not stable after 2 hours priming but Ill leave it on for another hour, maybe 2. I know it's supposed to be like 8 but 3~4 hours is a pretty good guideline what's stabel and what's not.

[Brother Esau]

Hmmm,

Ok guys, Ive some weird things going on.

I had some crashes while gaming. At most once a day though. I suspected my sound or videocards.

But as the ISP messed up once again yesterday I decided to have a look at the crashes and it's most likely from my RAM. So I played and played around. Drive Strength settings had the most effect. Timings (subtimings mainly) didnt hardly do anything, Voltage increase neither.

Prime95 ran at settings x 17~20 minutes before error, at settings y only 1 minute before error and settings z like BSOD. But there was nothing to make up out of it. Those settings were mainly between Drive Strength @ auto, something worked before and a mix from my succesfull DDR1140 settings.

So I left the Drive Strengths as I had them at the very first time, 1.0, 1.0, 1.0, 1.5, 1.5, 1.5, decreased the Vdimm by 0.05V and increased Vcore to 1.5375V. All subtimings on auto. Ran SuperPI 32M without problems, changed only MAL from 8 to 4 and tRC from 24 to 16 after that and ran it again without problems...

I didnt try Prime95 yet as I cant be arsed after 2 hours of BSOD's testing if it's fixed or not. But the fact it ran SuperPI 32M stable and even with decreased Vdimm... Not sure but Im pretty sure my RAM wouldn't run with lower Voltage before. That was quite suprising for me.

Also as it's 32M stable I think it's pretty close from being stable. Though stupid enough I didnt test with 32M when the BSOD's occured in Prime95.

Any idea though why at once it wasnt stable anymore? It's quite strange IMO. I dont think it's due to D9's, otherwise the decreased Voltage wouldnt have done anything I guess.

That was happening to me when the 1/30/08 Beta or 2/7/08 Beta corrupted my Bios chip if you flashed either of those 2 beta bios and have been having issues since I suggest you read back a couple of pages and see the Urgent Bios Info I posted regarding this situation!

[Rammsteiner]

Now you say... Im not entirely sure since when the problems started... I do have 1-30-08 BIOS, but I really cant remember if I had those problems before I went on vacation (22Feb~4March or so). I dont know either when I flashed the BIOS

Anyway, I read a few pages back as you suggested. And now you say the newest beta has a new flash utility which 'sees' the hole in the BIOS.

As Im not 100% sure what this really means; Ill download newest beta which seems to be stable according to you and others, do I still need all those parameters you suggested before for flashing back to 11/22 or does this 'fix' in the flash utility take care of this?

Also, if I drop HTT by 1Mhz, or change a few settings the system becomes more stable though. On the otherhand, the system already error'd in Prime95 after 4 hours and 50 minutes... That's pretty stupid IMO. Seems like there's no way to get it stable unless the Microns are really dead, but I dont want to believe that (yet).

Ill flash to newest beta as soon as I get a confirmation whether I should use the parameters with it or if it's not needed.

Thx in advance!

[Brother Esau]

Now you say... Im not entirely sure since when the problems started... I do have 1-30-08 BIOS, but I really cant remember if I had those problems before I went on vacation (22Feb~4March or so). I dont know either when I flashed the BIOS

Anyway, I read a few pages back as you suggested. And now you say the newest beta has a new flash utility which 'sees' the hole in the BIOS.

As Im not 100% sure what this really means; Ill download newest beta which seems to be stable according to you and others, do I still need all those parameters you suggested before for flashing back to 11/22 or does this 'fix' in the flash utility take care of this?

Also, if I drop HTT by 1Mhz, or change a few settings the system becomes more stable though. On the otherhand, the system already error'd in Prime95 after 4 hours and 50 minutes... That's pretty stupid IMO. Seems like there's no way to get it stable unless the Microns are really dead, but I dont want to believe that (yet).

Ill flash to newest beta as soon as I get a confirmation whether I should use the parameters with it or if it's not needed.

Thx in advance!

NO DO NOT ADD PARAMETERS! Just download 3/12/2008 Beta ZIP File and Boot from DRDOS and type AUTOEXEC.BAT if you continue to have issues do as I say in that long winded ordeal exactly please

[Rammsteiner]

Ok, thx.

This is why I love being drunk lol, Im extra suspicious for me to things up so Im patiently waiting for answers.

Anyway, Ill go flash it then. Just downloaded the newest beta.

But first, reconnect the floppy crap again *sigh*.

[Rammsteiner]

Crap, flashed, set everything to old settings, Prime and BSOD...

Looks like Ill have to mess around with those parameters then? If that doesnt work it will be most likely another problem then .

Anyway, how do I start from here regarding those parameters? Flash back to initial BIOS release and then back to this one?

[Brother Esau]

No its corrupted because you either flashed with the 1/30/08 or 2/7/08 Beta

Went through the same deal I would suggest you go back a few pages and read all of my posts and see if there is a similarity in you're situation compared to what I was posting. Do that first and get back to me with some feed back on your thoughts please and will advise at that point!

On a different not will be getting a Phenom 9600 BE so will help with that area with the Bios on this Board

[MKroenke]

Strange behaviour...I had both the 1/30 and the 3/7 beta, which did not really work properly but no issues with the latest beta. Why do you get corrupted BIOS chips?

[Brother Esau]

Why do you get a bad flash? Smame reason some how the Programming of the Flash EEProm does not go correctly. The issues with these Bios has been the way its been programmed during the flash procedure as well as other oversites from DFI.

[Oliverda]

No its corrupted because you either flashed with the 1/30/08 or 2/7/08 Beta

How can I check that my BIOS is corrupted or not? I always use the Award Winfash for flashing the BIOS. I like it.

[Brother Esau]

Stable settings that have been proven 8hrs + Prime stable on certain Bios for me it was 11/22/07 and irregardless of flashing to 11/22/07 3 times as well as other Bios it would still fail Prime within minutes at the exact same settings on the exact same bios that had been constant stable with those settings. After Flashing to and fro about 5 times I pulled my chip out of my board and put it on my Programmer and Erased/un scrambled it and programmed it with 11/7/07 and then a few hrs later at defaults I flashed to 11/22/07 Bios from DOS and proven stable settings that were stable but failed within minutes into prime 95 on 11/22/07 instantly became stable once again after Erasing & re-programming the chip on my programmer and updating Bios to 11/22/07 in DOS

So proven stable settings that have been tested extensively with a Certain Bios Release should yield the exact same result once you Flash back to that Bios if you update the Bios and find for what ever reason it does not work for you.
If that is not the case such as with me in my situation the only two conclusions I could make was Vibration damage to components from 2780 Mile Road trip in Big 26' Diesel Truck across country or Bios corruption

[justapost]

So, the 3/12/2008 BETA is the best BIOS for Phenom so far.



VALIDATION

Unfortunately it's unstable but it can boot up at least! I'm working on stable 3GHz right now.

I've found a strange bug in the AOD (both 0.14 and 0.17). It sets back the VCORE to the default 1.25V when I start it.

Hehe your 9500 rocks.
As for AOD you can modify the xml files in the program directory to force an other default for vcore. Or you can generate an profile with suitable settings and mark it for autoload.

[Brother Esau]

Hey, Achim whats up? Also good morning to ya if thats what it is where you are at anyways Notice nice sexy Phenom 9600 BE in siggy now?

[Oese]

assumed the bios was corrupted some way.. will simply flashing to new 03/12 beta fix this or would multiple flashing to initial release like you explained a few pages ago be the only way to recover?

I ask again because i'm not sure.. flashed from 01/30 to 02/07 beta relatively soon after getting the board, then to 03/07, then tried to recover but some sound issues made me flash the 01/30 again instead of 11/22, so before i flashed to 03/12 now i was messing with possible corrupted versions all the time and wondering whether this is the cause for relatively high vcore i need although under watercooling compared with other chips of the same stepping doing 3,2ghz at 1,38v easily...

[justapost]

...the one I am talking about as a option for entry level True Hardware Raid Controller from Areca is the Areca 1210 which is about $310.00 U.S Dollars and its a stellar super fast card I owned one before I bought this one and pound for pound its just as fast as this card minus the Burst Rate speed due to 256mb Onboard Ram

Hi Campbell,
I plan to buy such an controller for an linux server, have you tried the areca controllers under linux? If so any glitches?

Did you recieve your chips already? I'll finish my c2q testets till tomorrow and will continue testing with the dfi then. A backup bios chip whould be great.

achim~

[Brother Esau]

assumed the bios was corrupted some way.. will simply flashing to new 03/12 beta fix this or would multiple flashing to initial release like you explained a few pages ago be the only way to recover?

I ask again because i'm not sure.. flashed from 01/30 to 02/07 beta relatively soon after getting the board, then to 03/07, then tried to recover but some sound issues made me flash the 01/30 again instead of 11/22, so before i flashed to 03/12 now i was messing with possible corrupted versions all the time and wondering whether this is the cause for relatively high vcore i need although under watercooling compared with other chips of the same stepping doing 3,2ghz at 1,38v easily...

The Relevant Question here for you is have you ever had 8hrs+ Prime95 Stable on Blend Test? Have you ever Run MemTest 86 Over night without Errors? Have you ever Run OCCT CPU @ High Priority for 1hr stable? Have you ever dedicated that Kind of time on testing you're Overclock on any Particular Bios Release with complete success So you have a Basis on which you can judge any possible Future Bios Updates? If you failed My Test and answered NO then you should not be Flashing Bios Until you have a Reference Point to Compare with!

[Brother Esau]

Hi Campbell,
I plan to buy such an controller for an linux server, have you tried the areca controllers under linux? If so any glitches?

Did you recieve your chips already? I'll finish my c2q testets till tomorrow and will continue testing with the dfi then. A backup bios chip whould be great.

achim~

No I have not Run it Under Linux but I know allot that do and have and they say Areca and Linux is a winner!

The Bios Chips I have not ordered yet been really sick and in bed this past week because of the FLU So I will not see that shipment of chips for a couple of weeks and if you feel for what ever reason that You would feel safer or more comfortable getting one before I get mine in I want to encourage you to do what makes you feel the most comfortable

[justapost]

No I have not Run it Under Linux but I know allot that do and have and they say Areca and Linux is a winner!

Thx

The Bios Chips I have not ordered yet been really sick and in bed this past week because of the FLU So I will not see that shipment of chips for a couple of weeks and if you feel for what ever reason that You would feel safer or more comfortable getting one before I get mine in I want to encourage you to do what makes you feel the most comfortable

Was down with flu for three weeks during newyear, that was anoying (long). Found the chips for 12-15€ here guess i'll order two and try the live replace and flash method.
Do you think this method is save (replace the chips while under dos) or can I trash my chip easy that way?

[Oese]

@ be: of course i get your point. answer is yes (in part)! i tested prime blend and prime large fft for 6-8hrs each, but only with bios-versions that are possibly corrupt with respect to what you have stated. and i'm definitely stable, but with relatively high vcore. very very stable, not only with prime or something, but with overall system in game. never had any crashes since i own this system under these settings, and i can tell, this is definitely a fact that i appreciate very much about this board!!! every other overclocked system i had before was more troublesome to get it stable than this one.

i do not use memtest, but goldmemory, and there as well i got a thorough test (~8h) run without errors at my current settings. but are these the best possible settings really? or something crippled by possibly corrupted bios?

i would say no if you tell me that with a corrupted bios i would not get a stable system whatsoever at whatever vcore

and i would probably say yes if it could be possible to get stable settings with a corrupted bios on higher voltages

in fact, i wondered from the first day why i can boot into bios at 1,2v and into win win at 1,25v but need 1,45v (all at 3,2ghz) to get it stable running prime..

if this is possibly a sign for some problem related to what we are discussing, then: do i need to do the procedure you described again or would a dual-flash to 03/12 do the job as well (given some time left at defaults and the proper cmos clear..)?

i just want to confirm i run my board the best possible way as i got it second hand with 01/30 already installed...

[Brother Esau]

Thx

Was down with flu for three weeks during newyear, that was anoying (long). Found the chips for 12-15€ here guess i'll order two and try the live replace and flash method.
Do you think this method is save (replace the chips while under dos) or can I trash my chip easy that way?

You need PLCC Extraction Tool and there is a risk that when removing the Bios Chip or Replacing it can drop on the Motherboard and short it out or Damage the chip or both. So yes there is Risk involved not to mention ESD damage as well it should be done in a rea that is a ESD Work area or at the very least in the Kitchen or Room that does not have Carpeting and you should at a minimum have a ESD Wristband and don't go walking back and forth on Carpet and build up static electricity during the procedure matter of fact less walking or moving you do the better and make sure you have everything you need prior to doing this too.

[Oese]

I think I will try if i have time enough and tell whether theres any effect or not.. perhaps its simply my best possible overclock and thats it...

[Brother Esau]

@ be: of course i get your point. answer is yes (in part)! i tested prime blend and prime large fft for 6-8hrs each, but only with bios-versions that are possibly corrupt with respect to what you have stated. and i'm definitely stable, but with relatively high vcore. very very stable, not only with prime or something, but with overall system in game. never had any crashes since i own this system under these settings, and i can tell, this is definitely a fact that i appreciate very much about this board!!! every other overclocked system i had before was more troublesome to get it stable than this one.

i do not use memtest, but goldmemory, and there as well i got a thorough test (~8h) run without errors at my current settings. but are these the best possible settings really? or something crippled by possibly corrupted bios?

i would say no if you tell me that with a corrupted bios i would not get a stable system whatsoever at whatever vcore

and i would probably say yes if it could be possible to get stable settings with a corrupted bios on higher voltages

in fact, i wondered from the first day why i can boot into bios at 1,2v and into win win at 1,25v but need 1,45v (all at 3,2ghz) to get it stable running prime..

if this is possibly a sign for some problem related to what we are discussing, then: do i need to do the procedure you described again or would a dual-flash to 03/12 do the job as well (given some time left at defaults and the proper cmos clear..)?

i just want to confirm i run my board the best possible way as i got it second hand with 01/30 already installed...

A 4600+ @3.2GHZ is very good actually and I can pretty much tell you that you have a really good overclock on that chip considering what model it is What you should do at this point is see about lowering voltage and getting it stable that means 8hrs Prime Blend is what I judge by but thats me. Remember too that the more Voltage you give the Ram the more Vcore AMD cpu's require and sometimes it also requires getting the Hardware just right and that means buying stuff and then buying different stuff to take you're Overclock higher. but 3200mhz on a 4600+ is really good considering I can only get 3,6ghz out of my 6400+ BE maybe 3.7ghz on water tops thats only 400mhz Higher then Default clocks so all and all you're overclocking headroom is actually quite better then mine

But I will be selling my 6400+ BE if you may be interested????

[Oese]

@ be hehe thx!

the weird thing is i had a stable 6,5h prime at 1,4v after i freshly flashed 03/07 (i posted here if you remember) then next day i had a shutdown and after that no no way getting that stable again. nowhere near.

with regards to your 6400+: I would be interested but i think you want to have some money (and i suppose more than i would pay for buying a new one as it is pretested) and thats what i'm short of at the moment. and i'm thinking about moving towards phenom b3 and saving money for this step.. other option would be testing a 5200+, they are at 80€ in germany at the moment...

but if you wish you can make an offer pm-wise as of course a chip possibly doing a 3,6ghz oc watercooled would be VERY nice then we'll see