(omg)vflash | Fully Patched nvflash from X to Ada Lovelace [v5.780]

[zxTheWolfxz]

What card is this ?

Working with Tesla P100 wanting to flash QuadroGP100.
I'll post info from CMD window let me know if you think it is okay to type YES.
Edit: Didn't work.

Spoiler

C:\Users\GPU\Downloads\OMGVflash>omgvflash.exe --protectoff
NVIDIA Firmware Update Utility (Version 5.780.0)
Copyright (C) 1993-2022, NVIDIA Corporation. Veii Extreme Edition

Select a display adapter:
<0> Tesla P100-PCIE-16GB (10DE,15F8,10DE,118F) S:00,B:01,D:00,F:00
<1> NVIDIA TITAN X (Pascal) (10DE,1B00,1028,119A) S:00,B:02,D:00,F:00
Select a number (ESC to quit):
0
Setting EEPROM protection complete.

C:\Users\GPU\Downloads\OMGVflash>omgvflash.exe -6 --index=0 QuadroGP100.rom
NVIDIA Firmware Update Utility (Version 5.780.0)
Copyright (C) 1993-2022, NVIDIA Corporation. Veii Extreme Edition

Checking for matches between display adapter(s) and image(s)...

WARNING: None of the firmware image compatible PCI Device ID's
match the PCI Device ID of the adapter.
Adapter PCI Device ID: 15F8
Firmware image PCI Device ID: 15F0
WARNING: Firmware image PCI Subsystem ID (10DE.11C3)
does not match adapter PCI Subsystem ID (10DE.118F).
WARNING: None of the firmware image compatible Board ID's
match the Board ID of the adapter.
Adapter Board ID: EC3F
Firmware image Board ID: EC43

Downgrading/Sidegrading firwmare detected.
(This is usually only needed for very special circumstances.)

Are you sure you want to continue?
Type "YES" to confirm (all caps):

YES

Downgrade/Sidegrade is permitted.

GPU PCI Device ID override detected.
(This is usually only needed for very special circumstances.)

Are you sure you want to continue?
Type "YES" to confirm (all caps):
YES

Overriding the GPU PCI Device ID mismatch.

You are intending to override PCI Subsystem ID.
Are you sure you want to continue?
Press 'y' to confirm (any other key to abort): Y
Overriding the PCI Subsystem ID mismatch.

Board ID override detected.
*** WARNING: Overriding the Board ID can be very dangerous. ***
Upgrading to an image with the wrong Board ID might cause this system
to become or unstable.
Overriding the Board ID is only needed for extreme circumstances.
A mismatched Board ID almost always means the wrong firmware image is
being used for the specific switch.

Are you sure you want to continue?
Type "YES" to confirm (all caps):
YES

Overriding the Board ID mismatch.
Current - Version:86.00.4D.00.01 ID:10DE:15F8:10DE:118F
GP100 Board (Normal Board)
Replace with - Version:86.00.4D.00.04 ID:10DE:15F0:10DE:11C3
GP100 Board (Normal Board)

Update display adapter firmware?
Press 'y' to confirm (any other key to abort):

Nothing changed!



PROGRAMMING ERROR: EEPROM Program failed

C:\Users\GPU\Downloads\OMGVflash>

This is what I got with -L con.

Spoiler

C:\Users\GPU\Downloads\OMGVflash>OMGVflash -6 QuadroGP100.rom -L con

NVIDIA driver is not present

NVIDIA driver is not present
*** NVFLASH LOG FILE ***
Command-line : -6 QuadroGP100.rom -L con
NVIDIA Firmware Update Utility (Version 5.780.0)
Copyright (C) 1993-2022, NVIDIA Corporation. Veii Extreme Edition

NVIDIA Firmware Update Utility (Version 5.780.0)
Copyright (C) 1993-2022, NVIDIA Corporation. Veii Extreme Edition

Command: flash
FindDevices:
Device:0000:00:00:00=8086:2F00:8086:0000
Device:0000:00:01:00=8086:2F02:FFFF:FFFF
Device:0000:00:02:00=8086:2F04:FFFF:FFFF
Device:0000:00:03:00=8086:2F08:FFFF:FFFF
Device:0000:00:05:00=8086:2F28:0000:0000
Device:0000:00:05:01=8086:2F29:0000:0000
Device:0000:00:11:00=8086:8D7C:1043:8600
Device:0000:00:14:00=8086:8D31:1043:8600
Device:0000:00:16:00=8086:8D3A:1043:8600
Device:0000:00:19:00=8086:15A1:1043:85C4
Device:0000:00:1A:00=8086:8D2D:1043:8600
Device:0000:00:1C:00=8086:8D10:FFFF:FFFF
Device:0000:00:1C:04=8086:8D18:FFFF:FFFF
Device:0000:00:1D:00=8086:8D26:1043:8600
Device:0000:00:1F:00=8086:8D47:1043:8600
Device:0000:00:1F:02=8086:8D02:1043:8600
Device:0000:01:00:00=10DE:15F8:10DE:118F GPU
Device:0000:02:00:00=10DE:1B00:1028:119A GPU
Device:0000:02:00:01=10DE:10EF:1028:119A
Device:0000:05:00:00=1B21:1242:1043:8675
Filter-out on-motherboard PLX bridge device.
No PLX/PEX switch was detected, skip to filter-out non-NVIDIA PLX/PEX switch.
Checking for matches between display adapter(s) and image(s)...

Testing adapter: Tesla P100-PCIE-16GB (10DE,15F8,10DE,118F) S:00,B:01,D:00,F:00

Disable NV driver before start if NV driver loaded.

Adapter: Tesla P100-PCIE-16GB (10DE,15F8,10DE,118F) S:00,B:01,D:00,F:00

Identify EEPROM.
Action: Identify EEPROM.
uproc 05 ucode is not found for this GPU.
uproc 04 ready to be used.
Preparing secure operation after kicked start.
Actual SPI clock frequency (Hz): 16666667
Serial EEPROM identified.
EEPROM ID (EF,6013) : WBond W25Q40EW 1.65-1.95V 4096Kx1S, page

Action: Identify EEPROM.
Serial EEPROM identified.
EEPROM ID (EF,6013) : WBond W25Q40EW 1.65-1.95V 4096Kx1S, page

Action: Read EEPROM image.
Reading EEPROM bytes.
Start: 0x00000000
Count: 524288(0x00080000)
Read complete.=====================================] 100 %
Flash Status ledger is not present.

Testing Image: QuadroGP100.rom
IFR image mode
Invalid OTA signed VBIOS firmware image.
Current - Parsed Production VBIOS version: 86.00.4D.00.01
Replace with - Parsed Production VBIOS version: 86.00.4D.00.04

Mismatch Type: Computed: 0x00030207
WARNING: None of the firmware image compatible PCI Device ID's
WARNING: None of the firmware image compatible PCI Device ID's
match the PCI Device ID of the adapter.
match the PCI Device ID of the adapter.
Adapter PCI Device ID: 15F8
Adapter PCI Device ID: 15F8
Firmware image PCI Device ID: 15F0
Firmware image PCI Device ID: 15F0
WARNING: Firmware image PCI Subsystem ID (10DE.11C3)
WARNING: Firmware image PCI Subsystem ID (10DE.11C3)
does not match adapter PCI Subsystem ID (10DE.118F).
does not match adapter PCI Subsystem ID (10DE.118F).
WARNING: None of the firmware image compatible Board ID's
WARNING: None of the firmware image compatible Board ID's
match the Board ID of the adapter.
match the Board ID of the adapter.
Adapter Board ID: EC3F
Adapter Board ID: EC3F
Firmware image Board ID: EC43
Firmware image Board ID: EC43

Downgrading/Sidegrading firwmare detected.
(This is usually only needed for very special circumstances.)

Are you sure you want to continue?
Type "YES" to confirm (all caps):


Downgrading/Sidegrading firwmare detected.
(This is usually only needed for very special circumstances.)

Are you sure you want to continue?
Type "YES" to confirm (all caps):

YES


Downgrade/Sidegrade is permitted.
Downgrade/Sidegrade is permitted.

GPU PCI Device ID override detected.
(This is usually only needed for very special circumstances.)

Are you sure you want to continue?
Type "YES" to confirm (all caps):

GPU PCI Device ID override detected.
(This is usually only needed for very special circumstances.)

Are you sure you want to continue?
Type "YES" to confirm (all caps):
YES

Overriding the GPU PCI Device ID mismatch.
Overriding the GPU PCI Device ID mismatch.

You are intending to override PCI Subsystem ID.
Are you sure you want to continue?
Press 'y' to confirm (any other key to abort):
You are intending to override PCI Subsystem ID.
Are you sure you want to continue?
Press 'y' to confirm (any other key to abort): Y
Overriding the PCI Subsystem ID mismatch.
Overriding the PCI Subsystem ID mismatch.

Board ID override detected.
*** WARNING: Overriding the Board ID can be very dangerous. ***
Upgrading to an image with the wrong Board ID might cause this system
to become or unstable.
Overriding the Board ID is only needed for extreme circumstances.
A mismatched Board ID almost always means the wrong firmware image is
being used for the specific switch.

Are you sure you want to continue?
Type "YES" to confirm (all caps):

Board ID override detected.
*** WARNING: Overriding the Board ID can be very dangerous. ***
Upgrading to an image with the wrong Board ID might cause this system
to become or unstable.
Overriding the Board ID is only needed for extreme circumstances.
A mismatched Board ID almost always means the wrong firmware image is
being used for the specific switch.

Are you sure you want to continue?
Type "YES" to confirm (all caps):
YES


Overriding the Board ID mismatch.
Overriding the Board ID mismatch.
Mismatch Type: Adjusted: 0x00000000

Match found.
Current - Version:86.00.4D.00.01 ID:10DE:15F8:10DE:118F
GP100 Board (Normal Board)
Current - Version:86.00.4D.00.01 ID:10DE:15F8:10DE:118F
GP100 Board (Normal Board)
Replace with - Version:86.00.4D.00.04 ID:10DE:15F0:10DE:11C3
GP100 Board (Normal Board)
Replace with - Version:86.00.4D.00.04 ID:10DE:15F0:10DE:11C3
GP100 Board (Normal Board)

Update display adapter firmware?
Press 'y' to confirm ('s' to skip, 'a' to abort):
Action: Flash image.
Action: Identify EEPROM.
Serial EEPROM identified.
EEPROM ID (EF,6013) : WBond W25Q40EW 1.65-1.95V 4096Kx1S, page

Action: Read EEPROM image.
Reading EEPROM bytes.
Start: 0x00000000
Count: 524288(0x00080000)
Read complete.=====================================] 100 %
Flash Status ledger is not present.
H400.0500.01.04
H400.0500.01.04
InfoROM Release Level: 0x03(Current) v.s.0x03(File)
NOTES: GENERIC_DEFAULT = 0, GENERIC_POPULATED = 1
PRODUCT_DEFAULT = 2 ,PRODUCT_POPULATED = 3
InfoROM Release Version: H400.0500.01.04(Current) v.s. H400.0500.01.04(File)
INFOROM_FLASH_TYPE: 0x11
INFOROM_FLASH_TYPE_PRODUCT_ANY_TO_PRODUCT_POPULATED
Preserve OBD in IR block for default processing INFOROM_FLASH_TYPE_PRODUCT_ANY_TO_PRODUCT_POPULATED per absence of obdnopreserve.
Preserve IB for default processing INFOROM_FLASH_TYPE_PRODUCT_ANY_TO_PRODUCT_POPULATED per absence of obdnopreserve.
Preserve License image...
Preserve InfoROM Backup (IB) block from the source vbios in EEPROM.
Preserve InfoROM OBD object...
Succeed to preserve InfoROM OBD.
Preserve InfoROM OEM object...
Succeed to preserve InfoROM OEM object.
Preserve InfoROM PBL object...
Skip to preserve InfoROM PBL data for EEPROM does not have PBL data.
Final PCI expansion ROM size is 50688(0xC600) bytes
Appending InfoROM to the PCI ROM chain...
Final InfoROM Offset = 0x00037600 ( Absolute Offset 0x00038000)
InfoROM Offsets | ROMImage | InputAlt | Final |
NBSI OpRom Offset | 0x037400 | 0x000000 | 0x037400 |
NPDS Struct Offset | 0x037420 | 0x000020 | 0x037420 |
NPDE Offset | 0x037440 | 0x000040 | 0x037440 |
NBSI DIR Offset | 0x037FE0 | 0x000FE0 | 0x037FE0 |
JFFS Offset | 0x038000 | 0x001000 | 0x038000 |
Delete the current InfoROM block from the target vbios image.
Num of Objects = 10
Num of RW Objects = 6
Size of image with dummy RO blocks = 2816
Final PCI expansion ROM size is 50688(0xC600) bytes
Appending InfoROM to the PCI ROM chain...
Final InfoROM Offset = 0x00037600 ( Absolute Offset 0x00038000)
InfoROM Offsets | ROMImage | InputAlt | Final |
NBSI OpRom Offset | 0x037400 | 0x000000 | 0x037400 |
NPDS Struct Offset | 0x037420 | 0x000020 | 0x037420 |
NPDE Offset | 0x037440 | 0x000040 | 0x037440 |
NBSI DIR Offset | 0x037FE0 | 0x000FE0 | 0x037FE0 |
JFFS Offset | 0x038000 | 0x001000 | 0x038000 |
Num of Objects = 10
Num of RW Objects = 6
Size of image with dummy RO blocks = 2816
Num of Objects = 10
Num of RW Objects = 6
Size of image with dummy RO blocks = 2816
Num of Objects = 10
Num of RW Objects = 6
Size of image with dummy RO blocks = 2816
Skip proceeding with preservation table for Pascal or newer chip.
EEPROM size: 00080000
Image size: 00043A00
Unused size: 0003C600
Filled size: 0003C600

BCRT: Start Certificate 2.0 verification
Send VV Command...
Cert info block will be finalized during flash process.
Program page Start: 0x00000000 Count: 0x0A00(2560)
Command id: 0x30000005 Command: NV_UCODE_CMD_COMMAND_EWR failed
Command Status: NV_UCODE_CMD_STS_COMPLETE
Error Code = 0x0000005A(90): NV_UCODE_ERR_CODE_DEVID_MATCH_LIST_DEVID_MATCH_FAILED
Device ID in the rom image does not match the device ID in the chip.
Update aborted, please use a rom image that supports device ID 0x15F8.

Falcon CLOSE

EEPROM programming failed.

NOTE: Exception caught.
Nothing changed!
Nothing changed!






PROGRAMMING ERROR: EEPROM Program failed
PROGRAMMING ERROR: EEPROM Program failed
Exit Code: 4

C:\Users\GPU\Downloads\OMGVflash>

 

[Audioboxer]

Hey @Veii AB from Overclock here.

Thanks for your tool/work.

Am I correct to say with this BIOS VGA Bios Collection: GALAX RTX 4080 16 GB | TechPowerUp not only is the memory frequency permanently locked at full but the minimum voltage for core is also set at like 0.939v? As in, it won't go any lower.

I presume these are pre-baked settings for the HOF OC BIOS to help with extreme overclocking.

BIOS itself works great on my Gigabyte 4080, really hoping custom BIOS modding comes eventually so voltage can be unlocked alongside having a BIOS that normally downclocks memory/allows better undervolting curves. I've pulled up to 675w fine on water but it gets a bit hot lol. That's furmark for you.

Portal RTX can however easily pull 500W+ at 3440x1440 when clocked at 3.1ghz+ on core at 1.135v.

 

[_Purgatory_]

I have the Asus TUF RTX 4070 Ti non-oc. Can I use the Bios from the OC Version or from the MSI Gaming X Trio RTX 4070 Ti?

I also have a PNY 4070 Ti with 100% power limit. Can i flash a BIOS with a higher power limit ?

 

[Wotwow]

Tried to replace GOP for a newest version possible I could find for 3080 (GOP Update Tool does that automatically, much less fuzz then imhexing),

Give a link to your bios in which you need to update the GOP.

OK.
EVGA.RTX3080.10240.210313.rom (GOP 60009) to EVGA.RTX3080.10240.210313_0x60015.rom (GOP 60015)
Download:

UPLOAD.EE - File does not exist

File does not exist. Upload.ee

www.upload.ee

 

[SushYm4n]

Hey, many thanks to veii!
I just tried to flash a Gainward Bios onto my Inno3D RTX 4080 X3 for making possible a higher Power limit. Attached you see the result of my attempt.
btw. the result is indipendent of the commando, I also tried ./omgvflash.exe -6 Gainward340.rom
Any suggestions?

 

[Deleted member 218758]

This is what I got with -L con.

There is something left on nvflash's side , that i can and have to fix
CID rebrand is not functional for now (for quite a few cards already).
I'll push an update out when i got another Pascal series to play and verify
Approx 1-2 weeks , i hope.

Am I correct to say with this BIOS VGA Bios Collection: GALAX RTX 4080 16 GB | TechPowerUp not only is the memory frequency permanently locked at full but the minimum voltage for core is also set at like 0.939v? As in, it won't go any lower.

Would require dissasembly and a bios viewer. A bios viewer doesn't exist
Do not flash OC Lab bioses to normal cards unless you have a 2nd GPU to recover. Voltage controller compatibility issues.

Asus TUF RTX 4070 Ti non-oc

Yes. No MSI lightning tho.

PNY 4070 Ti with 100% power limit. Can i flash a BIOS with a higher power limit ?

No HOF , No EVGA Kingpin, No MSI Lightning, and no ASUS Matrix
Rest pretty much always works - if you find something interesting.
Be sure that at least one display port (HDMI/DP/TypeC) is in the correct location.
= Matches with yours. The rest adapts, mostly.

Any suggestions?

Needs a full rewrite,
Attach the -L log.txt nevertheless please
But due to EEPOM unlock of ISSI , different ones broke.
I need a full rewrite to fix that, but not an impossible task.
Use nvflashK for rebrands, if this one can't access your card

Will go for bug fixing soon.
Need gear that causes issues , to verify on it~

For my RTX 3080 Ti FTW3 Ultra is useless, right? It's already have got 450W cap (OC VBIOS).

Will respond within 24h
had XOC bios (not galax)

It looks like we currently are out of luck.
I asked a bit around but the answer is the same as from
"3080 Ti FTW3 cards are shipped with an "XOC" BIOS with a 450W power limit out of the box;"
They call 450W XOC https://forums.evga.com/EVGA-GeForce-30-Series-HYBRIDXOCDefault-BIOS-Thread-m3407542.aspx

One answer i'm still waiting for, but two were exactly that.
I guess there are Rebrands to turn it into a hybrid card and still retain BAR mode.
If there are other news, i'll let you know.
You could go and experiment with project-Hydra's curve-editor, if you feel like

Or potentially take a look if PrecisionX does allow control of FBVDD too.

 

[rusTORK]

If there are other news, i'll let you know.

It's fine. I know about their 450W XOC too. I thought maybe something i missed or something made for XOC overclockers, but it's pretty much GALAX 1000W BIOS for LN2. I am not THAT extreme. =)

3DMark hate me anyway. Solar Bay they accept result, but Port Royal - nope.

Thanks anyway! Focus on (omg)vflash then.

 

[PanosX]

you can flash 3080 to 3090 for unlocking cuda cores?is a same chip

 

[R-T-B]

you can flash 3080 to 3090 for unlocking cuda cores?is a same chip

No, laser cut dies.

 

[miguel80799]

Does anyone know how I can unlock the rgb of the 2060 super fe locked by bios? There are no bioses available for the 1f47 model to be able to change the color, the open rgb does not detect it and the evga software only allows it to be turned off and little else. translated by Google

 

[mrpaco]

has anyone managed to run dlss 3 on a 3000 series?

 

[Geek diy player]

Is it possible to bypass ERROR: PCI subsystem ID mismatch

 

[Deleted member 218758]

Is it possible to bypass ERROR: PCI subsystem ID mismatch

Please upload a -L log.txt
I read in OCN your question, but you said you edited the 3000 series ? bios.
It unlikely fails at pci-id error, it has to just said "file error" if you broke file-integrity.

Certification comes later ontop, but this is a smaller problem.

---

Currently i work on

Untitled - Pastebin.com

Pastebin.com is the number one paste tool since 2002. Pastebin is a website where you can store text online for a set period of time.

pastebin.com

Which basically is:
"EEPROM chipID rebrand , two stage"
then continue LL FW update through tool to complete. Quadro to Turing Consumer
Stuck at database-ID check from file to boardID (whitelist) , and stuck from rom CID check to database
Both can be fixed. This will take a while, but has a future.

It seems there is no done "start" of a 1000-4000 series biosviewer still.
No confirmed dev and no github project to contribute on. I hoped for more.

Past path above, CID & crossflash will be possible
If mix of SPI+Tool or Raw Tool access is possible, will have to be tested.
Falcon has playroom because it was never supposed to be patched against such shenanigans.
It boots up with manual CID rebrand & stays prepared for piggy-flashing, but of course SPI flash is not full and needs a stage two full-flash with omgvflash ~ to complete hardrebrand.
We'll see where we end up. I expect we wont need to dual stage flash for lack of CID permissions, but let's see~

I'll report back with PoW updates.

EDIT:
I still lack RL time to focus on this project more.
I hoped i won't be the only person working on this
Let's see what comes up. Pascal (and onwards) offset fixing from txt section is a correct approach , past 8bit blob fixing ~ but there hasn't been found a blob that is unsigned.
It doesn't boot yet, file integrity breaks.
Probably PPTable edit + 32bit fix, is the correct approach. But i try to edit blob so SID/CID can be edit or name of rom.
This requires double patching. Fixing blob checksum and then offsetting again from somewhere to restore expected file integrity. (this part still doesnt want to work)

* For mobile user inserted offset is a

But consumer doesnt allow to use the space inside a text field. Can't matroshka insert space to correct offset ~ in order to restore integrity.
We need another section that is unsigned, to do the file integrity offset fixing.
I haven't found a space that is unchecked yet.
GOP on desktop bothers with such shenanigans, although TDP Tweaker idea wasn't a bad one.
It just doesn't work.
And if it works, it has no future past 2000 series.
We need to take away our edits offset from somewhere else.

 

[Geek diy player]

How can i get the log.txt file？. Yes, I modified the maximum power consumption of the mobile 30 series, and I checked using the .\omgvflash.exe -A filename.rom command at the beginning of the article. Get PCI ID mismatch.。In fact I only changed one byte. Maybe broke something? Not sure, I'm not an expert on this.

 

[PanosX]

possible maybe 3080 10gb to flash 3080 12gb unlock 320bit to 384 and extra cuda cores

 

[Konstont1n]

Would require dissasembly and a bios viewer. A bios viewer doesn't exist

What about Ampere Bios Editor? https://github.com/bmgjet/Ampere-Bios-Editor
It can read some values from 3080/3090 early BIOSes. But the ability to save changes is locked.

It would be cool to have an editor like NiBiTot. With the ability to change memory timings and various power/voltage parameters.

This would be useful for hardware experiments.

For example, try to run new-fast Samsung memory from the 40xx card on a 30xx card and measure the performance changes.

Also there should be a function to disable memory channels in the BIOS. For quick and cheap software repair if video cards has problematic memory.

There is already a set of 30хх modified BIOSes with disabled memory channels from KrisFix-Germany.
But these are old versions of BIOSes and the set is not universal enough.

possible maybe 3080 10gb to flash 3080 12gb unlock 320bit to 384 and extra cuda cores

No. The characteristics of the chip, as a rule, do not depend on the BIOS.

 

[Deleted member 218758]

unlock 320bit to 384

320bit are trace design.
Not software design

What about Ampere Bios Editor? https://github.com/bmgjet/Ampere-Bios-Editor
It can read some values from 3080/3090 early BIOSes. But the ability to save changes is locked.

I have done the work they put, by hand. Know the layouts and couple more.
OCN helped user and worked together. Then dev sold the project to nvidia and is gone.
Its very limited. Doesnt support newer vbioses, doesn't support remain 3000 series stack.
Shows wrong checksum.
Let's not talk about it

But thank you for linking~

 

[PanosX]

Tried to replace GOP for a newest version possible I could find for 3080 (GOP Update Tool does that automatically, much less fuzz then imhexing),
Before that, running nvflashk.exe --version current.rom gave this:

Hierarchy ID          : Normal Board
Chip SKU              : 200-0
Project               : G132-0030
Build Date            : 10/23/20
Modification Date     : 03/13/21
UEFI Version          : 0x60009 ( x64 ) <-- VERSION BEFORE
UEFI Variant ID       : 0x000000000000000A ( GA1xx )
UEFI Signer(s)        : Microsoft Corporation UEFI CA 2011

after swapping the UEFI GOP module:

Hierarchy ID          : Normal Board
Chip SKU              : 200-0
Project               : G132-0030
Build Date            : 10/23/20
Modification Date     : 03/13/21
UEFI Version          : 0x60017 ( x64 ) <-- VERSION AFTER
UEFI Variant ID       : 0x000000000000000A ( GA1xx )
UEFI Signer(s)        : Microsoft Corporation UEFI CA 2011

However, nvflashk.exe cant flash it ...

C:\Users\zero\Downloads>nvflashk.exe -6 current_3080_0x60017.rom
nvflashk pre-release
github.com/notfromstatefarm/nvflashk - Safer GUI version with autorecovery coming by September!
Checking for matches between display adapter(s) and image(s)...
Reading EEPROM (this operation may take up to 30 seconds)
Current      - Version:94.02.42.C0.14 ID:10DE:2206:3842:3885
               GPU Board (Normal Board)
Replace with - Version:94.02.42.C0.14 ID:10DE:2206:3842:3885
               GPU Board (Normal Board)

Update display adapter firmware?
Press 'y' to confirm (any other key to abort):
Reading EEPROM (this operation may take up to 30 seconds)

Nothing changed!
ERROR: Invalid firmware image detected.

Too bad, would love to try this out finally

i have the same error likes you my original bios is 0x6009 goes to 0x60017 but failed

 

[zxTheWolfxz]

There is something left on nvflash's side , that i can and have to fix
CID rebrand is not functional for now (for quite a few cards already).
I'll push an update out when i got another Pascal series to play and verify
Approx 1-2 weeks , i hope

Thank you for your hard work on this looking forward to the next release.

 

[Deleted member 218758]

How can i get the log.txt file？. Yes, I modified the maximum power consumption of the mobile 30 series, and I checked using the .\omgvflash.exe -A filename.rom command at the beginning of the article. Get PCI ID mismatch.。In fact I only changed one byte. Maybe broke something? Not sure, I'm not an expert on this.

Please re'read the main post.
-A to identify the failure and flash with weak permissions, but community here has found -L log.txt, to be a better method for tracking
-6 romname.rom to flash the bios
-6 biosname.rom -L Log.txt , for debugging (upload the log here or on sites like pastebin & similar)

Thank you for your hard work on this looking forward to the next release.

(outer site emoji's like thumbs up don't seem to function on this forum)

 

[Geek diy player]

Hi Veii, Yes, at first I used -6 for testing and the program gave programming error: Failed to read EEPROM status register. Later I used -A to see the error and finally I uploaded the log attachment

EDIT: Is there any difference between this way and SPI refresh? I'll use the SPI programmer in my next teardown, give it a try.

 

[:D:D]

@Veii You seem to have spent a lot of time on this. There are no 32-bit checksums on Pascal and likely following VBIOS's so save yourself some time and enjoy some other things.

Using your nvflash or the other nvflashk does not flash a VBIOS that has been modded where it counts and signatures not updated. The "nvflashk - Flash any BIOS to NVIDIA GPUs" title is misleading in my IMO and should at least state "signed VBIOS".

As was seen in early Pascal days even flashing modded BIOS via SPI programmer would cause cards to not post except for mobile firmware. So what's different with mobile FW, embedded Falcon code?

Nvidia seems to have done a good job preventing modified VBIOS tweaks so unless a modded VBIOS can be re-signed, HULK cert broken (is there anyone actually using one of those?) or some back door found then there's little point IMO of developing a VBIOS editor/tweaker if the modded VBIOS cannot be used.

 

[Deleted member 218758]

Huh ? @:D:D
"You have spend a lot of time" - "you don't know, its not possible"
Decide which path you want to take~

nvflashK topic, in its corresponding thread.
I have nothing to do with it.

EDIT: Is there any difference between this way and SPI refresh?

Yes !

there's little point IMO of developing a VBIOS editor/tweaker if the modded VBIOS cannot be used.

I wish you would understand. But reading your text, you expect that nothing can be done.
// because the points you give , well half are due to past experience, other half are made up conclusions without ground.
People before you expect nothing can be done.
And we see 5 years of nothing doing.

Keep on with nothing doing
I disagree, but given to this page you haven't understood the intention.
I feel i have no idea how to explain it to you. You just can't relate and thats ok.

EDIT:
I'll get things myself done, like usually.

Hi Veii, Yes, at first I used -6 for testing and the program gave programming error: Failed to read EEPROM status register. Later I used -A to see the error and finally I uploaded the log attachment

Part of the file's signature, you tried to flash (3070 mobile)
is invalid.

Did you attempt to modify something or just grabbed one from techpowerup ?

Even if there is a potential to bypass - it's not really helpful, because it won't boot.
Going the path of breaking falcon is also not the right one to do.
Both are without a future, and even if potential backdoor exists and user reflashes falcon with proprietary software (can't)
It remains to have no future, as it will be patched in the very next bios.
Work with falcon not against it.

Using your nvflash or the other nvflashk does not flash a VBIOS that has been modded where it counts and signatures not updated

Correct. Pascal needs to have correct file integrity and onwards needs to have correct signature
But not "where it counts". This is your decision.
Where it counts is making correct bioses and have partition access to actually push it on ~ not expecting it to be easy like Maxwell and before.
Also purposely left enabled, because it makes no sense to let users brick their GPU with a ROM that will not boot.
If your file is made correct, falcon will allow it.

To add one thing more,
it can flash edited bioses, if the edit is correctly done. If falcon denies, the issue is on you not on the program.
I've verified this myself, it can flash edits ~ but nobody has said this is gonna be easy.

We can go and philosophe about past failed attempts of how impossible this is,
Or we use this thread space to contribute information about the process, in order to speed up the research
You decide~

 

[lerutis]

Just checking for fun, the 4070 ti :

Gigabyte RTX 4070 Ti VBIOS

12 GB GDDR6X, 2310 MHz GPU, 1313 MHz Memory

www.techpowerup.com

GPU Device Id: 0x10DE 0x2782 Version: 95.04.31.00.A3 GV-N407TEAGLE OC-12GD/F1/0E66 Copyright (C) 1996-2022 NVIDIA Corp. GPU Board Connectors 1x HDMI 3x DisplayPort Board power limit Target: 285.0 W Limit: 340.0 W Adj. Range: -65%, +19% Thermal Limits Rated: 84.0C Max: 88.0C Memory Support GDDR6X, Micron Boost Clock: 2625 MHz

Any change it will work on my normal 4070? Seems there is alot of similiar info, from what i can see though is that 4070 ti use a bigger pcie power connector, need more power for sure.

 

[zxTheWolfxz]

This is what I'm working with and only QuadroGP100.rom I could find at TPU.
I noticed my Tesla GP100.rom saved from the card is 271KB and the Quadro is 257KB.
Would the size make a difference when trying to flash and make it fail?
Wondering if it might need some padding.