Overclock.net › Forums › Graphics Cards › AMD/ATI › Software for R9 285 bios edit
New Posts  All Forums:Forum Nav:

Software for R9 285 bios edit - Page 52

post #511 of 1077
Sorry I have a mistake is usVddcOffset not usClockStretchAmounts. I'm going to change my previus messages. doh.gif . It's red highlighted B4FF-->-76, stock is CDFF-->-51 Warning: Spoiler! (Click to show)

Edited: I have changed 1250 in TongaBiosReader, and usVddcOffset with HxD.
Edited by mynm - 8/29/16 at 12:35pm
post #512 of 1077
Quote:
Originally Posted by mynm View Post

Sorry I have a mistake is usVddcOffset not usClockStretchAmounts I'm going to change my previus messages. doh.gif . It's red highlighted B4FF-->-76, stock is CDFF-->-51 Warning: Spoiler! (Click to show)
.

Would you be so kind to mod me a couple bios like you did the first but rather than 1250. Can you make one 1275, one 1300, 1325, 1350, 1375?

It's amazing what you did Mr Mynm, because it's totally hitch free running perfectly with your bios mod.

Haha I'd really appreciate it... Whatever you did makes it stable and low MVDDC volts. Without using Trixx at all, I just reset it upon start after AtiWinFlash. Which to flash in windows without fail start AtiWinFlash and end processes MSI, Trixx, HWinfo, RadeonSettings, and explorer and it succeeds every time.
post #513 of 1077
Quote:
Originally Posted by chris89 View Post

Would you be so kind to mod me a couple bios like you did the first but rather than 1250. Can you make one 1275, one 1300, 1325, 1350, 1375?

It's amazing what you did Mr Mynm, because it's totally hitch free running perfectly with your bios mod.

Haha I'd really appreciate it... Whatever you did makes it stable and low MVDDC volts. Without using Trixx at all, I just reset it upon start after AtiWinFlash. Which to flash in windows without fail start AtiWinFlash and end processes MSI, Trixx, HWinfo, RadeonSettings, and explorer and it succeeds every time.

Now I can't, try to do it yourself, with my info. If you can't I will do it if I have time.
post #514 of 1077
Quote:
Originally Posted by mynm View Post

Now I can't, try to do it yourself, with my info. If you can't I will do it if I have time.

Oh yeah AIDA64 doesn't show me any voltages for the gpu for some reason and I used the latest one. Your HWInfo shows more GPU info as far as Input voltage on the 12v rail etc mine doesn't show that. So sorry for not posting AIDA64 dumps, HWInfo is all I have.

I'm reading over your messages but it's a little difficult understanding in what offset. Atom bios reader says the location so I see in your 1.175vddci bios which is what im using shows the 0000A650 offset, Column 01 E2 02 04... I'm looking to up that 1250 value ..

http://www.overclock.net/content/type/61/id/2861316/width/500/height/1000/flags/LL

Thank you. You know the one value that was changed as seen in TongaBiosReader. Did you change that in hex or in Tonga? If I up the 1250, it auto results in very high mvddc but with your as-is just ups the core voltage and leave mvddc at a good low level. By the way I tested 1125 and it isn't stable, but 1094 is perfectly stable with your bios mod 1250 on the core which left mvddc at 1.188.

If you get time, maybe just 1 more mod. Just like before but instead of 1250, go 1350... I'll research on what you did in the mean time... So just open the .rom in HxD and search for the offset and alter the value basically binary to hex?
Edited by chris89 - 8/29/16 at 1:08pm
post #515 of 1077
Quote:
Originally Posted by mynm View Post

Warning: Spoiler! (Click to show)

I edited my BIOS to those values and my TimeSpy score was good with only 1.063 VDDCI. But at core clock 1195 I saw a type of screen corruption I have never seen before, so I actually increased VDDCI +25mV over stock and what happened is my VDDC voltage went up from 1.2v maximum to 1.225v.




chris89, updating HWINFO will change MVDDC to VDDCi... it was mislabeled. Everything can be changed by HEX editors alone; while tongabiosreader can only change some. gupsterg's 290/390 bios modding thread is a gold mine if your curious to learn.
Gaming Rig
(14 items)
 
  
CPUMotherboardGraphicsRAM
Intel 7700K ASRock Z270 K6 Gaming Fatality R9 285 TridentZ 3600 15-15-35 
Hard DriveCoolingOSMonitor
Samsung EVO 850 Corsair H115i Windows 10 Nixeus NX-Vue24a 
KeyboardPowerCaseMouse
Corsair K70 Rapidfire  Corsair RM850X Corsair 600Q Logitech G303 
Mouse PadAudio
Steelseries Qck+ Integrated  
  hide details  
Reply
Gaming Rig
(14 items)
 
  
CPUMotherboardGraphicsRAM
Intel 7700K ASRock Z270 K6 Gaming Fatality R9 285 TridentZ 3600 15-15-35 
Hard DriveCoolingOSMonitor
Samsung EVO 850 Corsair H115i Windows 10 Nixeus NX-Vue24a 
KeyboardPowerCaseMouse
Corsair K70 Rapidfire  Corsair RM850X Corsair 600Q Logitech G303 
Mouse PadAudio
Steelseries Qck+ Integrated  
  hide details  
Reply
post #516 of 1077
Quote:
Originally Posted by chris89 View Post

Oh yeah AIDA64 doesn't show me any voltages for the gpu for some reason and I used the latest one. Your HWInfo shows more GPU info as far as Input voltage on the 12v rail etc mine doesn't show that. So sorry for not posting AIDA64 dumps, HWInfo is all I have.

I'm reading over your messages but it's a little difficult understanding in what offset. Atom bios reader says the location so I see in your 1.175vddci bios which is what im using shows the 0000A650 offset, Column 01 E2 02 04... I'm looking to up that 1250 value ..

http://www.overclock.net/content/type/61/id/2861316/width/500/height/1000/flags/LL

Thank you. You know the one value that was changed as seen in TongaBiosReader. Did you change that in hex or in Tonga? If I up the 1250, it auto results in very high mvddc but with your as-is just ups the core voltage and leave mvddc at a good low level. By the way I tested 1125 and it isn't stable, but 1094 is perfectly stable with your bios mod 1250 on the core which left mvddc at 1.188.

If you get time, maybe just 1 more mod. Just like before but instead of 1250, go 1350... I'll research on what you did in the mean time... So just open the .rom in HxD and search for the offset and alter the value basically binary to hex?

If you can't use aida64, you can use hwinfo + gpu tweak 2.8.3.0 changing at 2d the values per DPM for core and memory and writing down values you see at hwinfo.

Can you do and i2c dump with aida64?
Quote:
Originally Posted by ducegt View Post

I edited my BIOS to those values and my TimeSpy score was good with only 1.063 VDDCI. But at core clock 1195 I saw a type of screen corruption I have never seen before, so I actually increased VDDCI +25mV over stock and what happened is my VDDC voltage went up from 1.2v maximum to 1.225v.




chris89, updating HWINFO will change MVDDC to VDDCi... it was mislabeled. Everything can be changed by HEX editors alone; while tongabiosreader can only change some. gupsterg's 290/390 bios modding thread is a gold mine if your curious to learn.

usVddcOffset have to be always negative and below 1.2v. 1.175v seems to be the max voltage for VDDCI with EVV voltages and it is achieve at 1010mhz, with 1.225v for core is 1.225 - 50 (ffcd usVddcOffset). So I prefer not to go over 1.175v for VDDCI. If you whant to go below 1.175v for VDDCI you have to edit usVddcOffset for lower DPMs, and that is what I did on the above image. I respected the diference between usVddcOffset, if DPM7 usVddcOffset is ffcd, and lower DPM are ffd8, to achieve a max of 1.1v for max DPM, I have changed it to ff69-->-151, so for DPM7 1.250v for core I will have 1.175v for VDDCI, and ff82-->-126 for lower DPMs.

Edited:The above I have stated is wrong you only have to change lower usVddcOffse it if you change manualy lower dpm voltages

Edited:
Quote:
Originally Posted by ducegt View Post

so I actually increased VDDCI +25mV over stock and what happened is my VDDC voltage went up from 1.2v maximum to 1.225v.

Are you talking about VDDCI instead of VDDC here? I have done "millions" of tests and I don't see that VDDC increase.
Edited by mynm - 9/2/16 at 4:41am
post #517 of 1077
Quote:
Originally Posted by mynm View Post

If you can't use aida64, you can use hwinfo + gpu tweak 2.8.3.0 changing at 2d the values per DPM for core and memory and writing down values you see at hwinfo.

Can you do and i2c dump with aida64?
usVddcOffset have to be always negative and below 1.2v. 1.175v seems to be the max voltage for VDDCI with EVV voltages and it is achieve at 1010mhz, with 1.225v for core is 1.225 - 50 (ffcd usVddcOffset). So I prefer not to go over 1.175v for VDDCI. If you whant to go below 1.175v for VDDCI you have to edit usVddcOffset for lower DPMs, and that is what I did on the above image. I respected the diference between usVddcOffset, if DPM7 usVddcOffset is ffcd, and lower DPM are ffd8, to achieve a max of 1.1v for max DPM, I have changed it to ff69-->-151, so for DPM7 1.250v for core I will have 1.175v for VDDCI, and ff82-->-126 for lower DPMs.

I need to download the latest HWInfo and AIDA64 to pull off a i2c dump haha give me some time, sure.

My card can handle high VDDCI probably no more than 1.325v realistically. It's the non-cooled VREGs which create instability. I use 17 watt per meter kelvin material on the memory modules, and core vregs. As the ram and aux vregs are not cooled and hoover at 80C all day internally thermal throttled, poor engineering choice. All vregs should and need to be cooled actively and directly for peak potential.

**I still need to get Windows 10... Do u know a good gui automated app that can setup a dual boot Win7/10 easily?

Back to topic yeah I downloaded AtomBiosReader-master, but it's a .jar like Java Gui and don't know how to open it? Do I need Linux or something? haha

I've been playing around and got 1250Mhz rocking and man it screams at 40 Giga Pixel's and 160 Giga Texel's but encounters anomalies around 1.350v, I think less than 1.35v is max maybe 1.325v.

I think it would be cool to set an offset in hex of the core voltage.

**Can you tell me once more what was done to that bios you sent me with 1.175 vddci?

You said actual is +50mv from that 1.175v so 1.225v which makes sense.... Can we leave the 65288, and just set an offset of +100mv to +125mv?

Your bios as-is 1.250v, runs great at 1094Mhz core and 1563Mhz memory. Here's my GPU-Z...

By the way I have (2) that's two AMD Radeon R9 290X 4GB models on my shelf, both shot but got for free. Baking now 450F 15-20min to correct them. I'm sure they would wax the R9 380X. haha


Edited by chris89 - 8/30/16 at 9:29am
post #518 of 1077
Quote:
Originally Posted by chris89 View Post

Warning: Spoiler! (Click to show)
I need to download the latest HWInfo and AIDA64 to pull off a i2c dump haha give me some time, sure.

My card can handle high VDDCI probably no more than 1.325v realistically. It's the non-cooled VREGs which create instability. I use 17 watt per meter kelvin material on the memory modules, and core vregs. As the ram and aux vregs are not cooled and hoover at 80C all day internally thermal throttled, poor engineering choice. All vregs should and need to be cooled actively and directly for peak potential.

**I still need to get Windows 10... Do u know a good gui automated app that can setup a dual boot Win7/10 easily?

Back to topic yeah I downloaded AtomBiosReader-master, but it's a .jar like Java Gui and don't know how to open it? Do I need Linux or something? haha

I've been playing around and got 1250Mhz rocking and man it screams at 40 Giga Pixel's and 160 Giga Texel's but encounters anomalies around 1.350v, I think less than 1.35v is max maybe 1.325v.

I think it would be cool to set an offset in hex of the core voltage.

**Can you tell me once more what was done to that bios you sent me with 1.175 vddci?

You said actual is +50mv from that 1.175v so 1.225v which makes sense.... Can we leave the 65288, and just set an offset of +100mv to +125mv?

Your bios as-is 1.250v, runs great at 1094Mhz core and 1563Mhz memory. Here's my GPU-Z...

By the way I have (2) that's two AMD Radeon R9 290X 4GB models on my shelf, both shot but got for free. Baking now 450F 15-20min to correct them. I'm sure they would wax the R9 380X. haha


usVddcOffset have to be always negative and VDDCI always below 1.2v. 1.325v VDDCI is to high will "burn" your gpu. VDDCI is the difference between 65288 (Edited:65288 is for DPM7 for the other DPMs are the respective 6528X value) and usVddcOffset, so usVddcOffset don't change VDDC only VDDCI.

I think that it is sufficiently well explained how I had mod the 1.175vddci bios Edited: If isn't, sorry, I can't explain it better. If you don't know how to do it, test your max vddc, core and memory clocks with Trixx or AB, and I will mod your bios.

AtomBiosReader .jar is at ReleaseJar folder.


Edited:

Acording to this gupsterg post, dynamic VDDCI can be disabled. I had tested to disable it, and it doesn't work for me, but I did the test via windows registry, not via bios. Perhaps via windows registry isn't working. Do somebody had tested it?.
Quote:
Originally Posted by gupsterg View Post

Tonga has "Dynamic VDDCI" IMO smile.gif .

Note: note the yellow background hex values (ignore rest as I'm working on PP).




Fiji is not "Dynamic VDDCI" IMO wink.gif .

I form this opinion from:-

a) Platform Caps info.

b) how Tonga has a usVddgfxoffset value in Tonga_MCLK_Dependency_Record and Fji does not.

c) usVddci and usMvdd are not equal in Tonga but are equal in Fiji.

d) your above test data (plus mine on Fiji).

You guys need to work on forming a relationship between the values in Tonga_MCLK_Dependency_Record if you wish to do edits to them.

I also think as Fiji does not have "Dynamic VDDCI" (perhaps due to the Platform Cap) when I set a usVddgfxoffset it caused freeze of GPU-Z render test (I also had usVddci & usMvdd equal).

Edited by mynm - 8/30/16 at 12:01pm
post #519 of 1077
Quote:
Originally Posted by mynm View Post

*usVddcOffset have to be always negative and VDDCI always below 1.2v. 1.325v VDDCI is to high will "burn" your gpu. VDDCI is the difference between 65288 (Edited:65288 is for DPM7 for the other DPMs are the respective 6528X value) and usVddcOffset, so usVddcOffset don't change VDDC only VDDCI.
*I think that it is sufficiently well explained how I had mod the 1.175vddci bios Edited: If isn't, sorry, I can't explain it better. If you don't know how to do it, test your max vddc, core and memory clocks with Trixx or AB, and I will mod your bios.
*AtomBiosReader .jar is at ReleaseJar folder.
*Acording to this gupsterg post, dynamic VDDCI can be disabled. I had tested to disable it, and it doesn't work for me, but I did the test via windows registry, not via bios. Perhaps via windows registry isn't working. Do somebody had tested it?.

Thanks mynm. I'm also a big time hex editing novice. On mobile gpu's in the past I select the start/end offset copy pasted edited via gui app, then pasted it back. Usually not editing directly, although I did once increase a 7150m gpu from 128MB dedicated to 1024MB which was neat. Anyway I am not familiar with .jar's and don't know how to open them, sadly.

I think I'd like to leave the 65288 value alone because MVDD rises too high when changing 65288 to 1250.

I found a +80mv (from previous bios mod 1250mv in dpm7) or 1.337v on HWInfo yielded stable 1,188Mhz though only because MVDD rose too high for 1219Mhz stable. Mvdd rises when upping volts on the core when 65288 is changed to 1250, which made mvdd rise to near 1.25v causing instability. With DPM7 at 65288, only setting a vgfx offset resulted in 1.35v core, and rock solid stability mvdd of just 1.188v.

Basically is the reason I'd like to set an internal offset of +125mv for about 1.35v core voltage leaving the dpm 7 value at 65288 because of MVDD rising too high.

**MY RESEARCH**
1063Mhz/1563Mhz stock 1.194v,1.188v
stock volts is 1.219v core, 1.188v ram
65288 stock bios +100mv 1,188Mhz yields 1.300v and 181 watts with memory no more than 1.188v at 11.222 watts.
Minimal TDP ceiling at 1,188mhz at 1.300v but 1,219Mhz at 1.300v hits the ceiling often.
So raising the TDP celing to 225 from 180 would open up 25% overhead since my thermals will certainly allow it.
At 225W TDP from 180, I could hit 1250Mhz
I found that 65288 bios allows 25mv less than with 65288 replaced with 1250. Meaning it can handle not only 1,188Mhz but also 1,219Mhz at 1.300v. (Referring to the 1.175 vddci bios mod)Where before it needed upwards of 1337mv just to keep 1,188Mhz stable and not 1,219Mhz.
1094Mhz needs 1244mv stable (+25mv offset)
1125Mhz needs 1263mv stable (+44mv offset)
1157Mhz needs 1275mv stable (+56mv offset)
1188Mhz needs 1337mv stable on 1.175 vddci bios mod, 1300mv on 65288 with +100mv offset
1219Mhz needs TDP increase, TDP won't allow it but 1.300v is possible
1250Mhz needs TDP increase, TDP won't allow it

Thank you very much sir.
Edited by chris89 - 8/30/16 at 12:58pm
post #520 of 1077
Quote:
Originally Posted by chris89 View Post

Warning: Spoiler! (Click to show)
Thanks mynm. I'm also a big time hex editing novice. On mobile gpu's in the past I select the start/end offset copy pasted edited via gui app, then pasted it back. Usually not editing directly, although I did once increase a 7150m gpu from 128MB dedicated to 1024MB which was neat. Anyway I am not familiar with .jar's and don't know how to open them, sadly.

I think I'd like to leave the 65288 value alone because MVDD rises too high when changing 65288 to 1250.

I found a +80mv (from previous bios mod 1250mv in dpm7) or 1.337v on HWInfo yielded stable 1,188Mhz though only because MVDD rose too high for 1219Mhz stable. Mvdd rises when upping volts on the core when 65288 is changed to 1250, which made mvdd rise to near 1.25v causing instability. With DPM7 at 65288, only setting a vgfx offset resulted in 1.35v core, and rock solid stability mvdd of just 1.188v.

Basically is the reason I'd like to set an internal offset of +125mv for about 1.35v core voltage leaving the dpm 7 value at 65288 because of MVDD rising too high.

Thank you very much sir.

Not a problem.thumb.gif


Do you have java installed?, perhaps you have an old not compatible verion. Is just click on AtomBiosReader.jar.

Yes, internal voltage offset will solve VDDCI rising, but I don't know how to do an internal voltage offset, I think that it can be done via BIOS VoltageObjectInfo table http://www.overclock.net/t/1561372/hawaii-bios-editing-290-290x-295x2-390-390x/2530#post_25102436 , but I don't have tested it. So the only way I know to lower VDDCI is via usVddgfxoffset.

And I see that not all 280, 380 and 380x have the same NCP81022 controller, I see one with an IR3567B controller: http://forums.guru3d.com/showpost.php?p=5222098&postcount=139 , so the mod is different depending the controller.

It will take my weeks to know how to do and test the internal voltage offset.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: AMD/ATI
Overclock.net › Forums › Graphics Cards › AMD/ATI › Software for R9 285 bios edit