Flashing full fat 680i bios to XFX 680i LT

[acidrain]

I own a full fat 680i motherboard in the form of an Asus P5N32-E SLI with an E6850 and OCZ PC-9200 reapers running at a 24/7 OC of 400Mhz FSB with DRAM in 1:1 mode on timings of 4-4-3-1T. I spent many hours getting this board stable (I can run it at a higher OC, but the 400MHz one is most efficient due to the asus fsb to northbridge strap that siginficantly loosens MCH timings above 400MHz and heavily penalizes memory bandwidth). During the process of OCing this mobo, I learned a lot about the 680i chipset - what it likes and what it doesn't.

So when I decided to put together a server machine, I bought the relatively cheap XFX 680i LT motherboard with an E8400 and a pair of the same OCZ PC-9200 reapers, thinking I'd get a respectable overclock. Well, I could OC the processor, with Orthos and Prime95 stable on a 3.6Ghz OC using a small FFT for 10 hours. To my dismay, however, as soon as I started stressing the CPU/RAM interface with a blend test, the system fell over. Usually within 3 minutes. (the 680i LT does not seem to like the 400Mhz FSB by the way - I wouldn't quite call it a hole, because it boots into Windoze, but it is very unstable).

I found on my Asus board that I needed to fix the MCP, CPU VTT, and HT voltages rather than letting the bios control them on auto settings in order to achieve a stable overclock. The 680i chipset does not like high voltages here. The trouble with the 680i LT, however, is that the crippled bios does not let me control these voltages. As a result, I could not run my RAM in synced and linked mode any higher than 633Mhz (i.e., a 1333MHz FSB) on very loose timings. I was getting a stock clock and no higher, and this annoyed me, as I knew the chip was capable of it and so was the memory. It was just the motherboard holding me back.

After fiddling for two days, I finally got impatient. I downloaded the XFX P32 BIOS for the full fat 680i. I then burned this BIOS, together with the 680i LT crippled P07 version to a bootable CD. I stuck the version of AWDFLASH.EXE shipped with my Asus driver disk onto the CD, as the XFX bios utility would not let me flash the P32 bios. I then booted the CD and, holding thumbs as I thought I had a good chance of bricking my mobo, flashed the P32 bios. All seemed to go smoothly. I rebooted, and lo and behold, a full 680i bios presented itself, with much finer grained control of voltages and other settings hidden in the 680i LT bios (such as some of the spread spectrum settings which default to auto and probably have a profound effect on any overclocking).

I tweaked the voltage settings. For reference purposes my voltage settings for a 3.6Ghz and a 4Ghz OC are, CPU VTT = 1.3v, SPP = 1.4v, MCP = 1.5v, HT = 1.25v. Everything disabled as per the nvidia 680i overclocking guide. I immediately tried to run the system at a 400Mhz FSB (3.6Ghz), but it was still not stable. Undeterred, I dropped the CPU multiplier from 9 to 8, and upped the FSB from 400Mhz to 450MHz, suspecting a FSB hole. Since the net effect of the multiplier drop and the FSB increase was to maintain the CPU clock at 3.6GHz, I knew that the CPU was stable (due to prior testing). Three hours later, the Orthos blend test was still chugging away merrily. I interrupted it, switched back into the bios, upped the CPU multiplier to 9, upped the vcore (I'd previously established that my CPU clock scales pretty linearly with vcore, requiring a 0.25mv vcore increase per 1Mhz of clock, so getting the right vcore was a fairly simple matter). The vcore for a 4050Mhz clock on my E8400 is 1.3875v, which is a little on the high side for my liking for a 24/7 stable overclock. Two hours later, Orthos was still chugging away on a blend test at 4050Mhz!!

The key to ANY form of overclock on this board for me was to control the voltages that the 680i LT bios would not let me control. I believe this incomprehensible limitation in the XFX 680i LT bios will hold back any reasonable overclocking of this board. As a result, I've uploaded the cd iso file to filefront.com for those who would like to try the same thing. The file can be found here (http://files.filefront.com/XFX+Custom+680iLT+BIOS+FLHzip/;9768332;/fileinfo.html). Download it, burn to a CD or DVD, boot off the CD. Then change to the R: drive where you will find the bios images and the flash utility. Execute one of the following

R:>AWDFLASH.EXE 691N9P32.BIN /e /py /sn /wb /cd /cp /cc /cks

To flash the full 680i P32 BIOS

OR

R:>AWDFLASH.EXE 721N9P07.BIN /e /py /sn /wb /cd /cp /cc /cks

To flash the crippled 680iLT P07 BIOS in case the previous flash aborts for some reason

WARNING: YOU DO THIS AT YOUR OWN RISK. THE MANUFACTURER RECOMMENDS AGAINST IT. YOU RUN A RISK OF BRICKING YOUR MOTHERBOARD, SO BE SURE THAT YOU ARE PREPARED FOR THIS EVENTUALITY. IT WORKED FOR ME, BUT THIS DOES NOT MEAN THAT IT WILL WORK FOR YOU.

If any of you do decide to do this, I'd be interested to hear of your experiences to see if they mirror my own.

Cheers,
AR

[Miker]

I would cut out your history and just say how to do it. A lot of people would like to learn how to put the BIOS on. Also, I wouldn't give the specs to what you set it to, just how you get the good BIOS in it.

Also, can we get a word from XFXSupport saying what this will do to the warranty?

[acidrain]

I though perhaps the voltage settings so critical to a stable overclock, as well as the identified FSB hole might be useful, hence the history. Sorry if it's a little verbose :wink: No doubt this will void my warranty, but since it was already voided by the installation of aftermarket chipset coolers, it's kinda academic to me..

[Miker]

People ask about this a lot, I am glad someone finally posted it. But I would turn it into a informational post.

As for the warranty, XFX is very forgiving. But, yes, chances are it voids your warranty.

[noobzed]

In the same way, possible to flash
xfx 790i with evga 790i bios ?

[hambone]

Anyone know where to find AWDFLASH exec to flash my XFX 680i LT from P07 to P31 (or 32) ??

I know it's out there somewhere but I'll be damned if I can find it.

[acidrain]

Anyone know where to find AWDFLASH exec to flash my XFX 680i LT from P07 to P31 (or 32) ??

I know it's out there somewhere but I'll be damned if I can find it.

That's what this whole thread is about. It's on the CD iso image linked at the top of this thread.

[spdracer]

After fiddling for two days, I finally got impatient. I downloaded the XFX P32 BIOS for the full fat 680i. I then burned this BIOS, together with the 680i LT crippled P07 version to a bootable CD. I stuck the version of AWDFLASH.EXE shipped with my Asus driver disk onto the CD, as the XFX bios utility would not let me flash the P32 bios. I then booted the CD and, holding thumbs as I thought I had a good chance of bricking my mobo, flashed the P32 bios. All seemed to go smoothly. I rebooted, and lo and behold, a full 680i bios presented itself, with much finer grained control of voltages and other settings hidden in the 680i LT bios (such as some of the spread spectrum settings which default to auto and probably have a profound effect on any overclocking).

I tweaked the voltage settings. For reference purposes my voltage settings for a 3.6Ghz and a 4Ghz OC are, CPU VTT = 1.3v, SPP = 1.4v, MCP = 1.5v, HT = 1.25v. Everything disabled as per the nvidia 680i overclocking guide. I immediately tried to run the system at a 400Mhz FSB (3.6Ghz), but it was still not stable. Undeterred, I dropped the CPU multiplier from 9 to 8, and upped the FSB from 400Mhz to 450MHz, suspecting a FSB hole. Since the net effect of the multiplier drop and the FSB increase was to maintain the CPU clock at 3.6GHz, I knew that the CPU was stable (due to prior testing). Three hours later, the Orthos blend test was still chugging away merrily. I interrupted it, switched back into the bios, upped the CPU multiplier to 9, upped the vcore (I'd previously established that my CPU clock scales pretty linearly with vcore, requiring a 0.25mv vcore increase per 1Mhz of clock, so getting the right vcore was a fairly simple matter). The vcore for a 4050Mhz clock on my E8400 is 1.3875v, which is a little on the high side for my liking for a 24/7 stable overclock. Two hours later, Orthos was still chugging away on a blend test at 4050Mhz!!

The key to ANY form of overclock on this board for me was to control the voltages that the 680i LT bios would not let me control. I believe this incomprehensible limitation in the XFX 680i LT bios will hold back any reasonable overclocking of this board. As a result, I've uploaded the cd iso file to filefront.com for those who would like to try the same thing. The file can be found here (http://files.filefront.com/XFX+Custom+680iLT+BIOS+FLHzip/;9768332;/fileinfo.html). Download it, burn to a CD or DVD, boot off the CD. Then change to the R: drive where you will find the bios images and the flash utility. Execute one of the following

R:>AWDFLASH.EXE 691N9P32.BIN /e /py /sn /wb /cd /cp /cc /cks

To flash the full 680i P32 BIOS

OR

R:>AWDFLASH.EXE 721N9P07.BIN /e /py /sn /wb /cd /cp /cc /cks



Cheers,
AR

Followed your method to a "T" on my xfx 680i lt and got " file does not match your part number. I heard there was something with the p32 that doesn't allow you to flash different boards,gonna try p31 and I'll post my results.

[boomstick68]

Has anyone else had success doing this? Also, R drive?? Are you meaning the CD/DVD drive? Mine is labeled D.

[acidrain]

The D: drive is the boot drive, the R: drive is where the bios images and awdflah utility live.

If you don't use the command line exactly as given, the flash will fail with the "file does not match your part number error". The flash needs to be forced.

[boomstick68]

LOL, yeah, I finally figured that one out. :beerchug:

[pablo NZ]

Hi acidrain,

Just like to thank you for the informative guide, the history and settings are also very useful _b

Just a question, what revision is your board? I have an A00 and it's very hard to find info on it, so just a little queasy about flashing the BIOS...

[pablo NZ]

update: ok so successfully flashed my board to the full fat bios, got lots more voltage options which is sweet, but it stills maxes out at 410MHz, exactly where it did before. Upping all the voltages didn't give me any extra headroom. :cry:

[acidrain]

I don't know what your machine setup looks like or your other bios options, but here are mine for a 4Ghz clock on my E8400 (http://www.overclock.net/intel-motherboards/304579-image-flash-full-p32-bios-xfx-2.html#post3647730)

[rumpleforeskin]

good guide there fella, i thought the history was interesting too as i have had pretty much the same experiance with this board as you had. I have been thinking of trying the 680 full BIOS for quite some time but was unsure if it would work or not.

Your post proved to me that it was possible so i have flashed the BIOS and it has booted with the full 680 goodness now available.

It remains to be seen wether these new options will let me get my chip back to 3.6 as on my intel board or if xfx mobo will still limit me to 2.8

thanks for being the 1st to test out the bios, as i would not have attempted it till i had enough cash for a replacement :)

[Halow]

First off, I want to apologize for bringing up an inactive and old thread, but after googling 30 different word variations, this is what I got.

Ok to my problem.

My Spec's Are...

XFX 680i LT Mobo
Intel Q6600 Quad Core Processor OC'd to 3.0
nVidia 7600 GS GFX card (updating soon)
2gb of OCZ PC6400 DDR2
600wat PW supply

I just flashed my XFX 680i LT mobo with the P32 iso provided in this thread -- with no problems whatsoever.

Before I had updated my BIOS, I OC'd my processor to 3.0 and it was working fine. But I kept getting errors from xp saying I needed to update my BIOS. I updated with the P32 BIOS because I wanted the extra voltage options.

After I updated the BIOS.. my processor is reading...

http://i27.tinypic.com/301hawy.jpg

As you can see, I increased my FSB to 1353 and its only reading a tad over 2.0. Does anyone have any information on this or anything that could help me? Is my processor really reading 2.0 when its 2.4 stock?

[bozotheclown]

it's pretty simple actually, you've probably got some of the powermanagement features in your bios enabled.
as you can see, your multiplier is 6 whilst is should be 9:)

When you overclock it's best to disable speedstep and c1e. If in addition to that, in the bios, your multiplier is set to auto or 9, your speed should be ok:)


First off, I want to apologize for bringing up an inactive and old thread, but after googling 30 different word variations, this is what I got.

Ok to my problem.

My Spec's Are...

XFX 680i LT Mobo
Intel Q6600 Quad Core Processor OC'd to 3.0
nVidia 7600 GS GFX card (updating soon)
2gb of OCZ PC6400 DDR2
600wat PW supply

I just flashed my XFX 680i LT mobo with the P32 iso provided in this thread -- with no problems whatsoever.

Before I had updated my BIOS, I OC'd my processor to 3.0 and it was working fine. But I kept getting errors from xp saying I needed to update my BIOS. I updated with the P32 BIOS because I wanted the extra voltage options.

After I updated the BIOS.. my processor is reading...

http://i27.tinypic.com/301hawy.jpg

As you can see, I increased my FSB to 1353 and its only reading a tad over 2.0. Does anyone have any information on this or anything that could help me? Is my processor really reading 2.0 when its 2.4 stock?

[tyle6]

agreed, turn off your power managements that are enabled by default in the full fat 680i bios. :smile:

[va3naa]

I wanted to extend my thanks to acidrain for starting this thread and providing the information on how to flash a 680i LT board with an 680i SLI bios. I followed the directions and it worked for me as well. Note that you must enter the line command he provided in uppercase or it won't work. I tried specifying the BIOS file name in lower case alpha characters and got a message saying the id's didn't match, but tried again with the BIOS file name using upper case characters and it worked well.

Note that when you flash your BIOS to P32, that you'll loose any previous settings you had and will have to respecify them. This is why one of the posts above noted that the default CPU multiplier dropped to 6 and that the speed step option was turned on again, yielding a clock speed less than the rated 2.4Ghz.

From another post, somebody wondered why the R: drive was specified in the line command, and they didn't have such a drive on their PC. If you look at the AUTOEXEC.BAT file in the ISO image, acidrain changes the first CD device on your computer to R:. This is temporary for the ISO boot and will revert back when you boot back into WIndows after the flash is complete.

Lots of extra voltage options available with the P32 BIOS installed. Good luck to those who venture down this path. A number of us have been successful following this procedure so you should be more comfortable with following the directions given and knowing you are likely to succeed.

Neil

[3991vhtes]

Will this work on a 630i board :lol:

[milton]

After downloading this modified file from the "The file can be found here." in the post and burning to CD, when I reboot the system the following errors appear.

CD-ROM DEVICE FRIVER FOR IDE (FOUR CHANNEL SUPPORTED)
OAK TECHMOLOGY
DRIVER V 340
DEVICE NAME BANANA
NO DRIVERS FOUND

DEVICE DRIVER NOT FOUND "BANANA"
NO VALID CDROM DEVICE DRIVERS SELECTED
A:\

Am I to modify the autoexe file? If so how?

I am running a internal LG LIGHT SCRIBE CD ROM SATA

Thx

[Mar10]

After downloading this modified file from the "The file can be found here." in the post and burning to CD, when I reboot the system the following errors appear.

CD-ROM DEVICE FRIVER FOR IDE (FOUR CHANNEL SUPPORTED)
OAK TECHMOLOGY
DRIVER V 340
DEVICE NAME BANANA
NO DRIVERS FOUND

DEVICE DRIVER NOT FOUND "BANANA"
NO VALID CDROM DEVICE DRIVERS SELECTED
A:\

Am I to modify the autoexe file? If so how?

I am running a internal LG LIGHT SCRIBE CD ROM SATA

Thx
Just plug in an IDE CD Rom till the bios Loads

[Mar10]

I own a full fat 680i motherboard in the form of an Asus P5N32-E SLI with an E6850 and OCZ PC-9200 reapers running at a 24/7 OC of 400Mhz FSB with DRAM in 1:1 mode on timings of 4-4-3-1T. I spent many hours getting this board stable (I can run it at a higher OC, but the 400MHz one is most efficient due to the asus fsb to northbridge strap that siginficantly loosens MCH timings above 400MHz and heavily penalizes memory bandwidth). During the process of OCing this mobo, I learned a lot about the 680i chipset - what it likes and what it doesn't.

So when I decided to put together a server machine, I bought the relatively cheap XFX 680i LT motherboard with an E8400 and a pair of the same OCZ PC-9200 reapers, thinking I'd get a respectable overclock. Well, I could OC the processor, with Orthos and Prime95 stable on a 3.6Ghz OC using a small FFT for 10 hours. To my dismay, however, as soon as I started stressing the CPU/RAM interface with a blend test, the system fell over. Usually within 3 minutes. (the 680i LT does not seem to like the 400Mhz FSB by the way - I wouldn't quite call it a hole, because it boots into Windoze, but it is very unstable).

I found on my Asus board that I needed to fix the MCP, CPU VTT, and HT voltages rather than letting the bios control them on auto settings in order to achieve a stable overclock. The 680i chipset does not like high voltages here. The trouble with the 680i LT, however, is that the crippled bios does not let me control these voltages. As a result, I could not run my RAM in synced and linked mode any higher than 633Mhz (i.e., a 1333MHz FSB) on very loose timings. I was getting a stock clock and no higher, and this annoyed me, as I knew the chip was capable of it and so was the memory. It was just the motherboard holding me back.

After fiddling for two days, I finally got impatient. I downloaded the XFX P32 BIOS for the full fat 680i. I then burned this BIOS, together with the 680i LT crippled P07 version to a bootable CD. I stuck the version of AWDFLASH.EXE shipped with my Asus driver disk onto the CD, as the XFX bios utility would not let me flash the P32 bios. I then booted the CD and, holding thumbs as I thought I had a good chance of bricking my mobo, flashed the P32 bios. All seemed to go smoothly. I rebooted, and lo and behold, a full 680i bios presented itself, with much finer grained control of voltages and other settings hidden in the 680i LT bios (such as some of the spread spectrum settings which default to auto and probably have a profound effect on any overclocking).

I tweaked the voltage settings. For reference purposes my voltage settings for a 3.6Ghz and a 4Ghz OC are, CPU VTT = 1.3v, SPP = 1.4v, MCP = 1.5v, HT = 1.25v. Everything disabled as per the nvidia 680i overclocking guide. I immediately tried to run the system at a 400Mhz FSB (3.6Ghz), but it was still not stable. Undeterred, I dropped the CPU multiplier from 9 to 8, and upped the FSB from 400Mhz to 450MHz, suspecting a FSB hole. Since the net effect of the multiplier drop and the FSB increase was to maintain the CPU clock at 3.6GHz, I knew that the CPU was stable (due to prior testing). Three hours later, the Orthos blend test was still chugging away merrily. I interrupted it, switched back into the bios, upped the CPU multiplier to 9, upped the vcore (I'd previously established that my CPU clock scales pretty linearly with vcore, requiring a 0.25mv vcore increase per 1Mhz of clock, so getting the right vcore was a fairly simple matter). The vcore for a 4050Mhz clock on my E8400 is 1.3875v, which is a little on the high side for my liking for a 24/7 stable overclock. Two hours later, Orthos was still chugging away on a blend test at 4050Mhz!!

The key to ANY form of overclock on this board for me was to control the voltages that the 680i LT bios would not let me control. I believe this incomprehensible limitation in the XFX 680i LT bios will hold back any reasonable overclocking of this board. As a result, I've uploaded the cd iso file to filefront.com for those who would like to try the same thing. The file can be found here (http://files.filefront.com/XFX+Custom+680iLT+BIOS+FLHzip/;9768332;/fileinfo.html). Download it, burn to a CD or DVD, boot off the CD. Then change to the R: drive where you will find the bios images and the flash utility. Execute one of the following

R:>AWDFLASH.EXE 691N9P32.BIN /e /py /sn /wb /cd /cp /cc /cks

To flash the full 680i P32 BIOS

OR

R:>AWDFLASH.EXE 721N9P07.BIN /e /py /sn /wb /cd /cp /cc /cks

To flash the crippled 680iLT P07 BIOS in case the previous flash aborts for some reason

WARNING: YOU DO THIS AT YOUR OWN RISK. THE MANUFACTURER RECOMMENDS AGAINST IT. YOU RUN A RISK OF BRICKING YOUR MOTHERBOARD, SO BE SURE THAT YOU ARE PREPARED FOR THIS EVENTUALITY. IT WORKED FOR ME, BUT THIS DOES NOT MEAN THAT IT WILL WORK FOR YOU.

If any of you do decide to do this, I'd be interested to hear of your experiences to see if they mirror my own.

Cheers,
AR
I have an OEM Board from Tiger Direct,It sill had the Nvida P05 in it! It went Right in,Now the P33 is out no need to Force it LOL.....Thanks

[spdracer]

I have an OEM Board from Tiger Direct,It sill had the Nvida P05 in it! It went Right in,Now the P33 is out no need to Force it LOL.....Thanks

I think the only version of the p33 is an EVGA. How do you propose to flash your xfx board?Thanks

[freya]

hey acidrain, will you make another file for flashing p33 bios for XFX 680i LT just like the current one but with p32? If not, then does anyone knows how to flash new bios onto the XFX 680i lt and if it works?:grin:

[freya]

i meant "but with p33"
sorry

[kramed]

hey acidrain, will you make another file for flashing p33 bios for XFX 680i LT just like the current one but with p32? If not, then does anyone knows how to flash new bios onto the XFX 680i lt and if it works?:grin:

I also would like to do this. I imagine you use the same command line parameters but download the P33 bios and use that instead.