VBE7 - vBIOS Editor for Radeon HD 7000 series cards

[lSD]

If I understand you correctly you want to set 1.200V as your voltage yes?

It's possible but that means I'll need to replace one of these voltages from the table so do you want to preserve 1.250V as max voltage or should I replace it with new voltage?

Of course I can replace any other voltage like 1.081V, but it's up to you.

well 1.250V as max voltage.
ps
sorry for my english

 

[KennyW]

Hey, thanks for reporting as it wasn't a display error in GPU-Z, but I actually forgot to change default voltage at one of the Offsets so your GPU in fact is still using 0.950V for 2D/UVD.

Anyway I've changed everything what was needed, so now all should be fine.

Works like a charm ! Thank you so much !

 

[Shambles1980]

dont suppose any one ever got round to editing my bios?
one thing i have noticed now.
is i could get away with only changing the lowest clock speeds to the second lowest ones.. (gou and ram) and also use whatever voltages that uses..
"just a matter of copy and paste the values and make sure the signing works?"

here is the initial post..
I dont need the 3d clock changed its fine at default, and after some experimenting the second state "some times gets used when playing flash videos and so on" is also fine.
the one that needs changing is the idle "lowest speed state" and that needs to be changed to exactly the same as the second lowest state (450mhz + whatever the ram speed is and the voltages) if yo need to know these values let me now and il try and find them for you. but they should be in the bios already as they are the same as the mid speed clock rate.

if you want to hex edit the bios that would be helpful.
just need it so all the clocks in all states are
900 gpu 1200 gddr preferably using 3d clocks.
don't need anything els changed really.

if you want to you could make a second edit with slightly higher voltages and 900/1200. but thats not really needed.
here is the rom file of the bios im using right now.

as the attatchment didnt come over with the quote here is the link to it
www.techpowerup.com/forums/attachments/pitcairn-zip.57196/

 

[sentye]

well 1.250V as max voltage.
ps
sorry for my english

OK, I've changed your voltage table slightly so you can set 1.200V as new voltage, although you will notice that there's no 1.081 voltage as I had to replace it.

Please test if everything is working properly and report back.

NOTE:
Before flashing new BIOS first make a backup of your original BIOS.

Regards

 

[sentye]

dont suppose any one ever got round to editing my bios?
one thing i have noticed now.
is i could get away with only changing the lowest clock speeds to the second lowest ones.. (gou and ram) and also use whatever voltages that uses..
"just a matter of copy and paste the values and make sure the signing works?"

here is the initial post..
I dont need the 3d clock changed its fine at default, and after some experimenting the second state "some times gets used when playing flash videos and so on" is also fine.
the one that needs changing is the idle "lowest speed state" and that needs to be changed to exactly the same as the second lowest state (450mhz + whatever the ram speed is and the voltages) if yo need to know these values let me now and il try and find them for you. but they should be in the bios already as they are the same as the mid speed clock rate.



as the attatchment didnt come over with the quote here is the link to it
www.techpowerup.com/forums/attachments/pitcairn-zip.57196/

Hi Shambles

Is my thinking correct and you want 2D state clocks to be the same as UVD?

Like this:

 

[Shambles1980]

Hi Shambles

Is my thinking correct and you want 2D state clocks to be the same as UVD?

Like this:

View attachment 57835

yes. it should stop the tearing i get at idle speeds unless i force 1080i res. or use dvxa or use a 3d heavy app..
its rediculous the card is fine in all other respects. but idle with drivers loaded produces what i can only describe are rediculous tearing issues.
heres a video i made for sapphire rma. which turned out to want stupid postage money so i decided to live with it..

as you can see it goes away when you start a 3d app of any kind. and later i found it also goes away if i force desktop to 1080i (30hz) games are fine though.. and just the other night i was forced to change to 800x600 to make a steam game work. and i noticed it wasnt doing it.. thats when i noticed flash vids stoped it and the clock rates were at 450.

(vid was taken before the rest of the system was updated but its the same problem)

tried everything els. 3d clocks are fine. and i just noticed the 450 speeds are fine too. so that would be better than setting everything to 3d clock speeds and voltages.
i would like a bios fix for it rather than my work arrounds. because when i upgrade my gpu i will pass this on to family or friends and they arent going to comprehend what i mean when i tell them how to work arround it. also 1080i is horrible lol the 30hz is death on the eyes

 

[sentye]

yes. it should stop the tearing i get at idle speeds unless i force 1080i res.
tried everything els. 3d clocks are fine. and i just noticed the 450 speeds are fine too. so that would be better than setting everything to 3d clock speeds and voltages.

That's a weird issue you have and it's a shame Sapphire didn't do anything about it as this could be fixed by a simple BIOS update. I've never experienced anything like this with my MSI 7850 even when I set 2D to 300/150 @ 0.800V everything was working fine, although I've seen other people reporting similar issues about flickering in 2D state with other GPUs, but most of them had been fixed by BIOS update.

I'm curious if it's a core/memory clock or voltage thing, because maybe only one frequency needs to be adjusted to fix such issue.

Anyway I did what you wanted so now hopefully everything will work as it should.

Regards

 

[Shambles1980]

thnx. its prety much the same issue that you can get on a 5770 if you over clock it, and the 2d clocks decide to run at stupid low speeds because of that. "well the same gfx issues at least" So i assume its a clock issue.
to oc a 5770 i had to use amd gpu tool which locked the card in to 3d states.. (nothing els was able to lock the cards state)
unfortunatly nothing works in regards to software clock alterations of the 2d clock or dissableing power play on this 7850. (but they didnt on the 5770 either) but now amd gpu tool wont work on the 7xxx cards so not many options available.
hopefully the bios will work. i wont get to test it out for a while as i go p[lay assetto corsa on monday evenings, so i wont flash the card untill after the games are over.

 

[Shambles1980]

! C:\Users\ShambleS\Downloads\Pitcairn_EDIT.zip: Unknown method in Pitcairn_EDIT.rom
! C:\Users\ShambleS\Downloads\Pitcairn_EDIT.zip: No files to extract


wont let me extract with above errors reported

edit.
7zip extracted it when nothing els would

 

[Marsellos]

! C:\Users\ShambleS\Downloads\Pitcairn_EDIT.zip: Unknown method in Pitcairn_EDIT.rom
! C:\Users\ShambleS\Downloads\Pitcairn_EDIT.zip: No files to extract


wont let me extract with above errors reported

WinRAR 5.10 or above

 

[Shambles1980]

was using win rar 7zip did it though

 

[Marsellos]

was using win rar 7zip did it though

Right now extracted perfectly WinRAR 5.10

 

[Shambles1980]

any way. just flashed the card the clocks are set like i asked for but the issue is still there which kinda sux.
i will just have to keep using 1080i when in 2d mode untill i spot something els that may fix it.

 

[hitem13]

Hello good sir,
What an awesome thread and great effort mister Thumbs up!
If its not to much trouble i would like to ask you something regarding ASUS HD7970 DCU2T 3GB card that i have.
I purchased a 120/144hz monitor and i have second monitor on 60hz, i had to flip the switch on my card for BIOS2 to get the 120/144hz support for my main monitor.
After a few days i notice the issue: My VRM1 temperature was hitting 90C in idle and even more during load, alittle to much for comfort.
GPU clock goes between 500 and 1000 while memory clock stays on 1400hz and powerconsumtions is maxed all the time, even in idle or when im surfing the webs.

Picture1: This is what i started with, default BIOS2 straight som powertech, no fan-profile.
Picture2: Added my own Fan-profile and managed to get it to 83C idle daytime and 77C idle nighttime (open window, 4:25am in sweden).

So i found this thread over at AMD and spoke some with a guy claiming GPU memory clock stuck att max when you have 2 monitor and one at above 60hz. Alltho i have tried everything and can not get my VRM1 temp to drop below 80C Idle (with my own fan profile: see attached bios). but it still reaches scary close to 100C during full load.
I have search for over 60 hours now for different solutions and suggestions and the more i read this thread the more i see some similarity's with the previous guys you have helped, when VDDC etc get stuck on a certain voltage. Maybe ive misunderstood the whole discussion you have had with them but hopefully ive made myself understandable.

Just to give some more information:
Ive been testing alot of configurations and it seems it dosent matter the amount of screens or what hz i run them in, with bios2 and 144hz support, the Voltage cranks to full and so does the memory clock (1400hz), no matter if idle, load or sleeping.
BIOS1 = Default bios. it has no 120hz/144hz support.
BIOS2 = From powertech w/144hz support

BIOS1
Monitors: 1x 60hz
VRM1 Temp Idle: Below 30C
VRM1 Temp Load: 56C after 2 hours

BIOS2
Monitors: 1x 120hz
VRM1 Temp Idle: 80-83C
VRM1 Temp Load: 95C+

BIOS2
Monitors: 1x 144hz - 1x 60hz
VRM1 Temp Idle: 80-83C
VRM1 Temp Load: 95C+

This makes me believe that it might not be related to the "2 screens, 1 above 60hz"-theory, unless its by design (i dont have the knowledge to tell how the bios is coded).
I think its the BIOS not scaling between power/memoryclocks vs load as this would give VRM1 plenty of time to cool off. The balancing feels broken.

Anyways, im running out of clues but they are pointing me in the direction of BIOS, so at this point any suggestions or any help would be greatly appreciated!
I have even orderd watercooling blocks and system for 300$ just to try to solve this.

Edit: GPU Tweak 2.1.7 shows 47C atm in idle on "power temprature" while HWINFO64 & GPU-Z both shows 80C. The temps correspond exactly so when +1C, they go to 48C vs 81C (so its a+33C difference on the sensor displays in the apps, maybe its gpu-tweak that is broken or other way around? hard to belive both hwinfo64 and cpu-z would randomly add +33C degrees to the displays?). Anyhow, added this info incase its usefull.

Kinds regards
Patrik

 

[scorpion1236]

Hi, how can I download this program: VBE7 ?

 

[Marsellos]

Hi, how can I download this program: VBE7 ?

First post of this topic:
http://www.techpowerup.com/forums/attachments/vbe7-0-0-7b-exe.52628/

 

[sentye]

any way. just flashed the card the clocks are set like i asked for but the issue is still there which kinda sux.
i will just have to keep using 1080i when in 2d mode untill i spot something els that may fix it.

Too bad it didn't work

However I've flashed your vBIOS over mine since they're almost the same to check if maybe I can reproduce the same issue on mine GPU, but it turned out that everything is working just fine as with my original BIOS.
I've tested it with my old monitor (DVI) and with HDTV (HDMI) and in neither of them I didn't saw any flickering or display corruption.

Have you tried flashing other BIOS to check if it may fix that issue, also does older drivers have any effect on that?

 

[Shambles1980]

yeah i have checked all sorts of bioses. its most likley a hardware error relating to 2d clocks only (not dxva/3d accelerated) any time the card goes in to a mode where its accelerating anything "videos or direct 3d even open gl" its fine.
it is strange how the issue goes away at 1080i "30fps" i dont really comprehend that part as the gpu isnt accelerating anything then.
but any thing from 800x600 - 1080p (60fps) when not being accelerated has that strange ass tearing. (the tearing is not there before drivers are installed though)

tried it on multiple systems and monitors all the same.
I have tried to force the card to always use 3d mode. but its still the same. so i just have to have my desktop at 1080i. but atleast games can play at correct resolutions.
It could be a lot worse. if the destop was fine and games were terrible then i would have spent the money to rma it.
But with postage costs, i could have sold this card as faulty. put the money i would have spent on postage towards a new one, and bought something like a r9 280. so it really wasnt worth it considering i had found a way to make it useable..

it could be a issue with powerplay. i cant seem to dissable it on this card. or force the card to use constant voltages.
the dxva acceleration is kind of a seperate entity in the gpu so it has its own acceleration. i had hoped setting the clocks to that speed would have meant the gpu would have decided not to tear, but i guess powerplay is still in affect even though the voltages are the same for no accelerated 2d and dxva acceleration.

the symptoms are exactly the same as when 2d clocks were to slow for a 5770. but the 2d clocks are fine.

i guess the only other option would be to leave dxva as it is. and set everything els to 860/1200 with 1.225v (i think its 1.225)but i dont know how powerplay would react to that.

 

[sentye]

yeah i have checked all sorts of bioses. its most likley a hardware error relating to 2d clocks only (not dxva/3d accelerated) any time the card goes in to a mode where its accelerating anything "videos or direct 3d even open gl" its fine.
it is strange how the issue goes away at 1080i "30fps" i dont really comprehend that part as the gpu isnt accelerating anything then.
but any thing from 800x600 - 1080p (60fps) when not being accelerated has that strange ass tearing. (the tearing is not there before drivers are installed though)

tried it on multiple systems and monitors all the same.
I have tried to force the card to always use 3d mode. but its still the same. so i just have to have my desktop at 1080i. but atleast games can play at correct resolutions.
It could be a lot worse. if the destop was fine and games were terrible then i would have spent the money to rma it.
But with postage costs, i could have sold this card as faulty. put the money i would have spent on postage towards a new one, and bought something like a r9 280. so it really wasnt worth it considering i had found a way to make it useable..

it could be a issue with powerplay. i cant seem to dissable it on this card. or force the card to use constant voltages.
the dxva acceleration is kind of a seperate entity in the gpu so it has its own acceleration. i had hoped setting the clocks to that speed would have meant the gpu would have decided not to tear, but i guess powerplay is still in affect even though the voltages are the same for no accelerated 2d and dxva acceleration.

the symptoms are exactly the same as when 2d clocks were to slow for a 5770. but the 2d clocks are fine.

i guess the only other option would be to leave dxva as it is. and set everything els to 860/1200 with 1.25v but i dont know how powerplay would react to that.

Some time ago I've set all state clocks and voltages the same as 3D (900/1200 @ 1.210) for testing and there were no issues, ofc apart from increased temperature. Also did the same what you wanted just left UVD as it is and I had zero issues with PowerPlay or with GPU itself.

If you want I can edit your vBIOS in such a way that it will always use max clocks and voltage for all states, except UVD, so you can test if this issue will be fixed.

 

[Shambles1980]

yes that would be helpfull

 

[sentye]

yes that would be helpfull

Done.

2D/3D clocks are the same, but UVD has been left untouched.

 

[Shambles1980]

just tried it but 2d clocks are still 450/1200 so not sure what happended there, i ran a you tube vid to test the dxva and thats 450/1200 so its not that the wrong ones were changed

 

[sentye]

Hello good sir,
What an awesome thread and great effort mister Thumbs up!
If its not to much trouble i would like to ask you something regarding ASUS HD7970 DCU2T 3GB card that i have.
I purchased a 120/144hz monitor and i have second monitor on 60hz, i had to flip the switch on my card for BIOS2 to get the 120/144hz support for my main monitor.
After a few days i notice the issue: My VRM1 temperature was hitting 90C in idle and even more during load, alittle to much for comfort.
GPU clock goes between 500 and 1000 while memory clock stays on 1400hz and powerconsumtions is maxed all the time, even in idle or when im surfing the webs.

Picture1: This is what i started with, default BIOS2 straight som powertech, no fan-profile.
Picture2: Added my own Fan-profile and managed to get it to 83C idle daytime and 77C idle nighttime (open window, 4:25am in sweden).

So i found this thread over at AMD and spoke some with a guy claiming GPU memory clock stuck att max when you have 2 monitor and one at above 60hz. Alltho i have tried everything and can not get my VRM1 temp to drop below 80C Idle (with my own fan profile: see attached bios). but it still reaches scary close to 100C during full load.
I have search for over 60 hours now for different solutions and suggestions and the more i read this thread the more i see some similarity's with the previous guys you have helped, when VDDC etc get stuck on a certain voltage. Maybe ive misunderstood the whole discussion you have had with them but hopefully ive made myself understandable.

Just to give some more information:
Ive been testing alot of configurations and it seems it dosent matter the amount of screens or what hz i run them in, with bios2 and 144hz support, the Voltage cranks to full and so does the memory clock (1400hz), no matter if idle, load or sleeping.
BIOS1 = Default bios. it has no 120hz/144hz support.
BIOS2 = From powertech w/144hz support

BIOS1
Monitors: 1x 60hz
VRM1 Temp Idle: Below 30C
VRM1 Temp Load: 56C after 2 hours

BIOS2
Monitors: 1x 120hz
VRM1 Temp Idle: 80-83C
VRM1 Temp Load: 95C+

BIOS2
Monitors: 1x 144hz - 1x 60hz
VRM1 Temp Idle: 80-83C
VRM1 Temp Load: 95C+

This makes me believe that it might not be related to the "2 screens, 1 above 60hz"-theory, unless its by design (i dont have the knowledge to tell how the bios is coded).
I think its the BIOS not scaling between power/memoryclocks vs load as this would give VRM1 plenty of time to cool off. The balancing feels broken.

Anyways, im running out of clues but they are pointing me in the direction of BIOS, so at this point any suggestions or any help would be greatly appreciated!
I have even orderd watercooling blocks and system for 300$ just to try to solve this.

Edit: GPU Tweak 2.1.7 shows 47C atm in idle on "power temprature" while HWINFO64 & GPU-Z both shows 80C. The temps correspond exactly so when +1C, they go to 48C vs 81C (so its a+43C difference on the sensor displays in the apps, maybe its gpu-tweak that is broken or other way around? hard to belive both hwinfo64 and cpu-z would randomly add +33C degrees to the displays?). Anyhow, added this info incase its usefull.

Kinds regards
Patrik

Hello hitem

Unfortunately this what's happening with your card is by design and it's not only AMD, but also Nvidia cards which are affected by this, they did something which will force higher clocks to prevent instabilities when using multiple monitors, or even a single monitor with refresh rate set higher than 60Hz (although for 120Hz it should still use idle clocks). I'm not sure if there's anything in vBIOS related to that what's happening in such case or if maybe there's something they set in drivers, nevertheless I'll see if I can find anything in your vBIOS which will make your GPU to use lower clocks for this matter.

Just one question, have you tried different inputs i.e. DVI/HDMI/DP and does it change anything?

Regards

 

[sentye]

just tried it but 2d clocks are still 450/1200 so not sure what happended there, i ran a you tube vid to test the dxva and thats 450/1200 so its not that the wrong ones were changed

Well I've checked that BIOS and everything seems fine to me.

Did you used "Extend official overclocking limits" option in MSI AB?
If you're unsure: run regedit, find following reg key - PP_PhmSoftPowerPlayTable, if it's found delete it and restart your PC.

I've found that it may interfere with actual reading of 2D/UVD clocks and tools like GPU-Z will report frequencies from that reg key.

 

[Shambles1980]

i use trix and dont need to extend anything with that. but i do have msi installed so i will have a search for it.

uninnstalled msi and restarted and the clocks changed to what was set in bios..
but problem is still the same.
im gonna just blame power play and forget about it ithink..

thanx for the edits though. atleast i know it cant be fixed with just clock and voltage rates.