DFI users - want a GOOD BIOS???

Printable View

Show 100 post(s) from this thread on one page

05-23-2004, 10:16 AM
STEvil

Quote:

Originally posted by AdvModDev
Out of curiosity, what problems were you having with Oskar's original 4/29 that made you flash to EvilouS' 429 C18?

I never tried oskar's 429, went straight to 1/31 from the stock 11/27.

Quote:

Just so you know, the way 1/31 is coded (tigher timings at the 200MHZ ROMSIP table), it should technically outperform the 12/18. I have the same problem in that my config doesn't agree with 1/31, but it was always faster than 1/21 or 12/18 (these two have nearly identical ROMSIPs, BTW). But these DFI mobos are so totally unpredictable and picky.

You got that right :D

Quote:

Tictac and the Abit BIOS modders are essentially doing this as they are taking the 12/18 ROMSIP and inserting into their tweaked AN7 and NF7 BIOSes.

Yes, but do their bioses also open up all the extra timings these infinity/lpb bioses have? We already know we can use Epox 8rda+'s for hotflashes, so maybe it may be doable to just flash some of these other boards directly to infinity/lpb bioses and have all these extra little timings at our disposal without having to modify bioses for every single board out there.

Also, i'll give up a bit of performance for stability, but I like to have every option out there available to me in case i should get the sick desire to go for broke... not that that doesnt happen a lot... :eek: ;) :toast:
05-23-2004, 10:17 AM
AdvModDev

Quote:

Originally posted by SAE

At Adv:

I use 6F = 9F and 6E = 01 with wpcredit... it's what's in the thread over at AMDMB. It works better for me, more stable and the 1F setting is actually set by every bios :confused: This is the setting which caused the weird noises many users reported. 9F makes it go away ;)

And I could up the clocks more stably in windows...

I know SAE, Tromoto is aware that the latest Oskar Betas (at least) have direct coding of register values in the BPL module. The patch works (ie sets 6C to 9F) with the official 11/27 BIOS for sure. You're right about the default setting of 1F, but remember the default is the "old value": 1F0FFF01. NV has recommended a new timeout value of: 1F01FF01. This is what the patch does for the latest Betas (at least). I wasn't aware that the "choke buzzing" was due to 1F, at least it doesn't affect my machine. Tromoto did say that some Abit users that have implemented this patch have indeed been able to reach higher stable overclocks. Like I said before, he's still trying to work out the 1F, 9F thing but at least the patch gives the correct timeout value (even though 1F is intended for the older revision chipsets).
05-23-2004, 10:44 AM
WildKard

I use 9F/01 like SAE, 1F sometimes I crash/dont load windows, 9F seems infinately stable, but every system is different of course

I have managed to finish my testing

505 T6 12/18

Speed = 2nd
Stability = 3rd
Peripheral drop = 3rd

505 T6 1/21
Speed = 3rd
Stability = 1st*
Peripheral Drop = 1st*

505 T6 1/31
Speed = 1st
Stability = 1st*
Peripheral Drop = 1st*

*Tied

Thats about 4 days of bios testing, gaming/priming/music listening/general computer usage. I experienced no lockups/crashes with the 1/31 or 1/21 bios romsip versions, 2 hardlocks with 12/18

All testing was done using 255fsb x 11 @ 2811mhz 11-2-2-2-2.0 EEFD 1.8vdd 1.75vcore 3vdimm

Of course every system will vary and if you recieve these bios from anybody flash at your own risk, they are NOT OFFICIAL.

Thanks to all the hardwork that went into this, as now I finally have a newer bios version that works for my system and performs great :)

*EDIT* Alpha timings are 1-3-3-3-3-3-4 for maximum speed/stability
05-23-2004, 10:48 AM
AdvModDev

Quote:

Originally posted by STEvil
I never tried oskar's 429, went straight to 1/31 from the stock 11/27.

So you went from 11/27 -> 1/31 -> EvilouS' modded 4/29?

Maybe it's just me, but it just don't make no sense to not test stability/performace of the original (Oskar's 4/29) before taking the risk of testing any modded BIOS of the same version. In my case, I had stability problems with the original 4/29 and only after realizing this did I even think about modding it.

Quote:

We already know we can use Epox 8rda+'s for hotflashes, so maybe it may be doable to just flash some of these other boards directly to infinity/lpb bioses and have all these extra little timings at our disposal without having to modify bioses for every single board out there.

LOL, that's pretty nucking futs. One example: the NF7 uses the 3112 SATA chip, which doesn't support RAID 1+0 or 4 channel like the Infinity/LPB's 3114 chip. If you tried to use a 3112 SATA BIOS integrated into the NF7 BIOS on a DFI, you would have some serious execution problems at the least. :D
05-23-2004, 10:56 AM
AdvModDev

Quote:

Originally posted by WildKard
I use 9F/01 like SAE, 1F sometimes I crash/dont load windows, 9F seems infinately stable, but every system is different of course

I have managed to finish my testing

505 T6 12/18

Speed = 2nd
Stability = 3rd
Peripheral drop = 3rd

505 T6 1/21
Speed = 3rd
Stability = 1st*
Peripheral Drop = 1st*

505 T6 1/31
Speed = 1st
Stability = 1st*
Peripheral Drop = 1st*

*Tied

Thats about 4 days of bios testing, gaming/priming/music listening/general computer usage. I experienced no lockups/crashes with the 1/31 or 1/21 bios romsip versions, 2 hardlocks with 12/18

All testing was done using 255fsb x 11 @ 2811mhz 11-2-2-2-2.0 EEFD 1.8vdd 1.75vcore 3vdimm

Of course every system will vary and if you recieve these bios from anybody flash at your own risk, they are NOT OFFICIAL.

Thanks to all the hardwork that went into this, as now I finally have a newer bios version that works for my system and performs great :)

*EDIT* Alpha timings are 1-3-3-3-3-3-4 for maximum speed/stability

That's totally awesome WildKard :) That is strange that the 1/31 version didn't give you any drops (maybe my methodology is indeed flawed :D ) If it was 100% right, then you should've gotten the same results as you would with the original 1/21, 12/18, and 1/31, respectively. Sincere thank-you for the significant thoroughness of your testing and of course, your time. I'm satisfied that whatever work I did helped you in any way, brother WildKard.
05-23-2004, 08:05 PM
STEvil

Quote:

Originally posted by AdvModDev So you went from 11/27 -> 1/31 -> EvilouS' modded 4/29?

Maybe it's just me, but it just don't make no sense to not test stability/performace of the original (Oskar's 4/29) before taking the risk of testing any modded BIOS of the same version. In my case, I had stability problems with the original 4/29 and only after realizing this did I even think about modding it.

I dont like flashing bioses as i'd rather not have a bad flash.. havent had one yet, but....

Quote:

LOL, that's pretty nucking futs. One example: the NF7 uses the 3112 SATA chip, which doesn't support RAID 1+0 or 4 channel like the Infinity/LPB's 3114 chip. If you tried to use a 3112 SATA BIOS integrated into the NF7 BIOS on a DFI, you would have some serious execution problems at the least. :D

Indeed, but what about nF7 non S? Or other boards... does SATA even matter if the board is just for benching? Maybe the SATA part of the bios can be changed out easier than the many romsips that are changed or need to be changed the current way?

I'm not disputing that running one board with a bios that was not meant for it might be a bad idea, but it may be worth a try.. :stick:
05-23-2004, 09:49 PM
DisposableHero

Quote:

Originally posted by WildKard
I use 9F/01 like SAE, 1F sometimes I crash/dont load windows, 9F seems infinately stable, but every system is different of course

I have managed to finish my testing

505 T6 12/18

Speed = 2nd
Stability = 3rd
Peripheral drop = 3rd

505 T6 1/21
Speed = 3rd
Stability = 1st*
Peripheral Drop = 1st*

505 T6 1/31
Speed = 1st
Stability = 1st*
Peripheral Drop = 1st*

*Tied

Thats about 4 days of bios testing, gaming/priming/music listening/general computer usage. I experienced no lockups/crashes with the 1/31 or 1/21 bios romsip versions, 2 hardlocks with 12/18

All testing was done using 255fsb x 11 @ 2811mhz 11-2-2-2-2.0 EEFD 1.8vdd 1.75vcore 3vdimm

Of course every system will vary and if you recieve these bios from anybody flash at your own risk, they are NOT OFFICIAL.

Thanks to all the hardwork that went into this, as now I finally have a newer bios version that works for my system and performs great :)

*EDIT* Alpha timings are 1-3-3-3-3-3-4 for maximum speed/stability

whered you get the modded 1-21 ??
05-23-2004, 09:56 PM
megatron

AdModDev is only relasing via email but Evilous has them freely available:
http://forums.pcper.com/showthread.php?t=321614
05-23-2004, 11:51 PM
V I P E R

Hello guys,
I have tried the 5/5 modded BIOS from Evilous and the BIOS corrupted after 3 days. I have flashed my chip with *.b02 which worked OK with my setup, but when I try to power my PC only the LED 1 and 2 are working and my system doen't want to boot ... I can't get into BIOS. In the manual LED 1 and 2 mean that the system is initialising the Onboard Super IO. Can you help me and tell me how to solve this problem. I did not changed anythig exept the BIOS and now my system is dead.
05-24-2004, 12:03 AM
STEvil

remove the cmos battery, wait 20 seconds, try powering on without the battery.
05-24-2004, 12:24 AM
V I P E R

Quote:

Originally posted by STEvil
remove the cmos battery, wait 20 seconds, try powering on without the battery.

OK I'll try that, but what should I expect when I try that?
I have tried to clear CMOS several times but nothing changed.
05-24-2004, 04:54 AM
megatron

Then hot flash it. There are tricks like holding insert or 1 key on power up.
05-24-2004, 08:08 AM
G|-|oST

@ VIPER

Did you see the big red writing at the bottom of the Evilous thread? Your scenario is exactly what the warning is about.

Elsewhere in the same thread there are several recovery techniques described by several people, because after all the warnings etc. people still flash without knowing what they are getting themselves into.

These BIOSes are not for everyone! :stick:

edit: spelling
05-24-2004, 12:23 PM
Naasar

alright, now we are cooking. i got the 5/5 bios from someone on pc perspective forums. BAAM!!!!, memory is normal and i can boot from my 80 wd drive......

my goodness, i never thought my ram would hit a wall .....lol

trying to stablize 255 fsb @ 2,2,2,2,11:banana: :toast:

thanks everyone for the help.

along came dfi:eek:
05-24-2004, 01:58 PM
FlyingHamster

let us know your settings if u hit 255fsb stable plz. :)
05-24-2004, 02:11 PM
WildKard

I dont think 255fsb is my boards limit, I think its my memories limit

I was running a P4 3.0C @ 255fsb and I hit the same wall(8-4-4-2.5) that I hit on the infinity @ 11-2-2-2-2.0 I just can never pass 255, very odd :) I can run memtest @ 255 without errors but as soon as I hit 256+ it errors out instantly its like it was designed to crap out at this speed
05-24-2004, 03:15 PM
STEvil

Can I get a copy of the bios you guys are running?

I'm gonna wait until I have my hands on a bios savior or extra bios chip, but it'd be nice to have a copy until then.
05-24-2004, 03:16 PM
SAE

Quote:

Originally posted by WildKard
I dont think 255fsb is my boards limit, I think its my memories limit

I was running a P4 3.0C @ 255fsb and I hit the same wall(8-4-4-2.5) that I hit on the infinity @ 11-2-2-2-2.0 I just can never pass 255, very odd :) I can run memtest @ 255 without errors but as soon as I hit 256+ it errors out instantly its like it was designed to crap out at this speed

I just thought like that with my Infinities and my HyperX as I could not get anything stable in dualchannel over 240... but could bench at 260 sometimes... weird I thought and wondered if new mem would kick a$$... but Then, as I got my LP B all probs went away, I could immediately do dual channel at 260MHz with the same ram... memtest stable!!!

Go figure ;)
05-24-2004, 05:26 PM
AdvModDev

Quote:

Originally posted by STEvil
I dont like flashing bioses as i'd rather not have a bad flash.. havent had one yet, but....

Wanting to avoid flashing as much as possible (especially if one doesn't have a backup plan), now that makes sense to me.

What doesn't make sense to me is the guy who flashes a modded 4/29 or 5/5 directly without first giving the original Oskar 4/29 or 5/5 a chance. If after testing out the performance/stability of the originals (and one finds out that performance/stability is not up to par with, say, the 1/21), then it would make sense to think about flashing a modded version.

Naaser is a perfect example of someone who did the above. He flashed an EvilouS modded 5/5 with x ROMSIP (correct me if I'm wrong) and his config's reaction was below par. Then he flashed the original Oskar 5/5 and his expectations were met. He could've saved alot of time and hassle if he took the sensible route in the first place.
05-24-2004, 09:23 PM
Rolle2k

Now i have tested the Three versions with 1/21 ROMSIP. the 6 Table version was the only one i could boot over 240FSB with, i didn't get it really stable as i had with the original 1/21 Bios, but with the original bios i had problems loosing sound somtimes if i went a bit over 220FSB, i have experienced NO problems with that on this bios. Will try the 6 Table version with 1/31 romsip later today. Be back later with more results when i have tested it. im at work right now. Anyway great work AmdModDev!

and also, if Toromoto comes up with a patch for the Gigalan as you said before.. will you include that one in the bioses?
05-24-2004, 09:31 PM
WildKard

After extended periods of use, I have discovered I lose lan at high FSB with 1/31 romsip, other than that it is AOK!

So, I now use 505 T6 1/31 with a nic :) This is way faster than using the old old 1/21 bios and I finally have a system that runs the way I want it to(and the way it should!)

Thanks for all your help modding the bios :)
05-24-2004, 09:40 PM
V I P E R

Quote:

Originally posted by G|-|oST
@ VIPER

Did you see the big red writing at the bottom of the Evilous thread? Your scenario is exactly what the warning is about.

Elsewhere in the same thread there are several recovery techniques described by several people, because after all the warnings etc. people still flash without knowing what they are getting themselves into.

These BIOSes are not for everyone! :stick:

edit: spelling

I'm not somebody who doesn't know what he is doing.
I have a friend with professional chip programmer and he flashed my BIOS chip several times, but nothig changed. I stil get the same LED 1 and 2 flashing and I can not get in BIOS. My friend saved my a$$ lots of times because I have flashed my BIOS many times and the flash with the programmer always helped but now nothing helps. I have tested my hardware on other systems and everything is OK.
I have tried all recovery techniques but nothing changed. I think I'll RMA the board because this modded BIOS fu**ed my mobo.
05-24-2004, 10:16 PM
STEvil

Why does it have to make sense? I wanted to flash the 429-12/18 evilous and did so.. thats good enough for me :/
05-24-2004, 10:35 PM
EvilouS

Quote:

Originally posted by AdvModDev
Well, I see you guys want to turn this thing into somekind of competition. :rolleyes:

I can't believe I'm even responding to this, but think about this for a little bit: what are we trying to accomplish here? We are trying to replace the less stable 5/5 ROMSIP with the more stable 1/21 ROMSIP. If the methodology used is correct (ie if the ROMSIP from x BIOS was truly inserted into x BIOS) then, logically, performance and stability should be identical.

But this really can't be done in this case because the 1/21 ROMSIP and the 5/5 ROMSIP are not structured the same. Not only do they have different structures, they differ in total size (the 1/21 ROMSIP has 10 Tables and the 5/5 ROMSIP has 16 Tables). So to have a 5/5 BIOS w/ a genuine 1/21 ROMSIP is both logically and physically impossible.

If Oskar had followed his usual 10 Table approach for the 5/5 BIOS, then modding it for real would be a 3min job, if that. There would be no doubt if the methodology is correct and there would be no need for lengthy testing whatsoever. But this, unfortunately, is not the case. I sincerely hope EvilouS understands this.

Now, you guys have forced me to go back to AMDMB/PCP DFI forum (aka FLAME FORUM) to download the file you guys did for the comparison. Actually I was just curious:D, but I've stopped going there for real because while it used to be a good forum, it has deteriorated significantly. Anyways, the comparison done by Megatron is 1) for two different BIOSes and 2) totally different ROMSIPs. EvilouS' uses the 4/29 BIOS, while mine uses the 5/5 BIOS. I have used a methodology that I believe is closest to the actual 1/21 ROMSIP. His methodolgy for the ROMSIP done in the comparison does not resemble the 1/21 ROMSIP at all. Actually, I suspect that he has used Tables from an ABIT BIOS for that one.

Jesus, I can't believe that I actually feel the need to explain any of this to anyone.

Look, if you guys read my post above you know where I stand about all of this BIOS modding and what my motives were/are. By all means, if you have BIOS saviors and know what you are doing, go to AMDMB and try all of EvilouS' modded BIOSes. He is a good guy and wants to help people. I wish him the best of luck and hope he figures it all out successfully. I also hope he understands his repsponsibilities for accepting such a burden. Maybe he can forget about the one guy that accidentally kills his board flashing an EvilouS modded BIOS. But, you see, I cannot. 'Nuff said.

Hi Guys EvilouS here

I’ve just read this whole thread and would like to say a few things

Firstly I’d like to thank AdvModDev for his work, coz he has put a lot of effort into it.
I can’t speak for him but I enjoy doing it as I’m addicted to overclocking like so much of you. So anytime I mod a bios for myself I’m gonna let all of you have it as well, but please don’t expect schedules and miracles, and remember they have been created for MYSELF.
The ROMSIPS on each of the bioses differ in timings and lengths, so you can’t take a ABIT ROMSIP and replace the LPB ROMSIP with it for example however you can take a couple of the tables and experiment with it. Which is what I did with the SPC bios. It is in no way competing with AdvMod’s superb 121 versions, it’s not even meant to be that fast because the timings for this bios are slower than that of the DFI so it was meant to allow higher clocks. Any modified bios will beat this clock for clock. Therefore it will be and will always remain an experiment.

All my newer bioses come with Tromoto’s HALT fix so there’s no need to ask for them. And yes it is in the NV NIC bios so you need it enabled for it to work.

Like AdvModDev I don’t like hearing people fry their boards when flashing my latest recipe that is why I have a big fat red disclaimer and warning at the bottom of the thread, if you can’t flash correctly don’t flash at all.
We’ve now lost two fantastic modders TICTAC and now AdvModDev because the burden is too much. It’s a shame coz their work is so great, and more of a shame coz I would love to have done collaboration with AdvModDev

Keep overclocking’
05-24-2004, 11:12 PM
megatron

Hey Evilous it was nice of u to allow AdModDev to use ur romsips that u extracted manually yourself, since as u said this is the most time consuming part. I don't know why but he got very upset when I posted a comparison which someone else asked for and I thought why not. He even pm'ed me about this. I also never released the bioses he sent me at his request. Seems to me that he wants to steal your limelight.

Show 100 post(s) from this thread on one page