Question GIGABYTE 5700 XT Bios mod fails

[AlleyCat]

Hi,
I am failing to update Gigabyte 5700 XT. The same procedure updates the BIOS on MSI cards with no problem. I follow the instructions on Igors Lab.

The sign of having trouble with the card bios flash shows in GPU-Z. After flash using amdvbflash the values of GPU and memory frequencies is empty. With stock bios there are Mhz frequencies.

Is there any known problems with flashing Gigabyte cards?
Any suggestions on what other forums I may ask for assistance?
Is it possible that the OEM bios is signed, and any modified bios will be rejected?
If the bios is signed, any tools to resign, or would I need to buy a card from a different vendor?

Thanks,

Alley Cat

 

[Mini_Me]

has anyone worked with the MSI MECH OC 5700 xt? I have apple straps applied and am working on the other settings now.

What do you need?

 

[AlleyCat]

I can actually tell you as a fact that changing mem clock from 905 to 910 has minimal impact (at least with 780/850/1350) on those powercolor 5700XT.
About 0.2-0.3 MH/s.
Changing from 1420 to 1400 will take of a chunk off (about 0.5 MHs).
The drop im talking about is coming from lowering the core frequency (the 1380/1385). Actually everything below 1440 has impact, but the jump from 1420 to 1440 takes a good amount of wattage for similar gains like the 1400 to 1420 change.
Thats what made me end up with 1420 in the first place.

I could probably get the same hashrate with 1420 core and 910 or 905 plus your mem voltages.

I will try when the fork of ETC has cleared up

I found that the optimal clock to memory speed is 1410MHz, with 780mv, and running the memory at 930MHz with 850mv.

I get 57MH/s and stability. The 1410 clock speed helps the 930 memory to pull some hash without increasing the power significantly.

If I get a card dead error, I increase the VDD by 10mv, or reduce the memory clock by 10MHz.

 

[dan]

I found that the optimal clock to memory speed is 1410MHz, with 780mv, and running the memory at 930MHz with 850mv.

I get 57MH/s and stability. The 1410 clock speed helps the 930 memory to pull some hash without increasing the power significantly.

If I get a card dead error, I increase the VDD by 10mv, or reduce the memory clock by 10MHz.

Like mentioned before i run 1420 780/850/1350 910 applied apple timings and no changes to powertables.
Regular Powercolors 5700XT.
On fake MH/s rate for phoenix i get 56,9 MH/s for samsung and 56,8 MH/s for micron.

"If I get a card dead error, I increase the VDD by 10mv, or reduce the memory clock by 10MHz."
I do the same but i go 5MHz steps on memory

 

[bolebol]

hello. i have ASUS Dual Radeon™ RX 5700 EVO OC edition 8GB GDDR6. and uploadet xt bios on it whit apple inc straps. and wonder if you have some input on best settings in hiveos for make it run stable clock 24/7

 

[Sartori71]

Once again I would like to thank all the contributors to this thread for the time and effort put in to testing. I do have my rig stable for the last couple of weeks. Now that I have a bit of time I would like to dial it in a little more. For those that are using Gigabyte 5700xt Gaming OC cards, can you post your voltage and memory settings? There is a lot of information in this thread, so I’m just looking for some final settings that you’ve come up with. I’m using the apple straps and MPT bios mod settings from this thread, and my hiveOS settings are below. I have a difference in wattage reported by 2 of my cards as they are the rev 2 of the same card, and thanks to another contributor to this thread it looks like that is just a software issue, actual wattage at the wall appears to be the same, so thanks for that as well!

 

[aposhtol]

has anyone worked with the MSI MECH OC 5700 xt? I have apple straps applied and am working on the other settings now.

Windows, 20.11.2 drivers, Apple straps, 56.5Mh/s (Phoenix miner), power consumption (as reported by software) <=90W


Core 1310MHz (miners report 1280Mhz)
Memory 1808Mhz (miners report 1800Mhz)
cvvd 710mV

set with AMD Memory Tweak XL:
mvvd 1325mV
vddci 675mV
SoC 950Mhz

 

[Mini_Me]

Once again I would like to thank all the contributors to this thread for the time and effort put in to testing. I do have my rig stable for the last couple of weeks. Now that I have a bit of time I would like to dial it in a little more. For those that are using Gigabyte 5700xt Gaming OC cards, can you post your voltage and memory settings? There is a lot of information in this thread, so I’m just looking for some final settings that you’ve come up with. I’m using the apple straps and MPT bios mod settings from this thread, and my hiveOS settings are below. I have a difference in wattage reported by 2 of my cards as they are the rev 2 of the same card, and thanks to another contributor to this thread it looks like that is just a software issue, actual wattage at the wall appears to be the same, so thanks for that as well!

Greetings,

I might figured out a way to bypass the rsa signature of the new vbios of the revision 2,
Please reflash the original F3 vbios then modify its power play table with morepowertool and apply the vram timings and save it then send it to me,
Please do apply the power table in the page below,

Question - GIGABYTE 5700 XT Bios mod fails

Hi, I am failing to update Gigabyte 5700 XT. The same procedure updates the BIOS on MSI cards with no problem. I follow the instructions on Igors Lab. The sign of having trouble with the card bios flash shows in GPU-Z. After flash using amdvbflash the values of GPU and memory frequencies is...

www.igorslab.de

 

[montecarlo]

Hi, just come across this thread, I purposely took out 2 of my card out to test this "Apple Strap". Doesn't like my card. I'm still in search of the lowest power 5700xt. I've seen people doing it 55~60 @ 105-115W (from the wall). I believed only certain binned cards with mod bios can do that.

Do note that software reporting power is nothing close to the wall. In average, a tuned 5700 xt runs at 56~60 @ 135~145w (95~105 in software) from the wall. Software power reporting will also varies if you flash different bios into it or altered the 'Power and Voltage' part of the MPT.

 

[AndSer]

I got an additional question if you dont mind.
Are those RX 5700 XT Mech OC or regular RX 5700 XT Mechs?
I would like to give MSI a try, you would help me out to choose because i can get my hands on either of them.
The OCs are the ones i see too much heat problems written about, but i would give it a try anyway if you got those.
Thx

I'm sorry for late response.
I have RX 5700 XT Mech OC on all rigs now. GPUs in one rig are based on Micron memory, all the others - Samsung. I set them (all rigs) not to exceed 57MH/s as at the higher hash revenue ratio suffers and GPU temp goes above 82C. Naming convention from Looney Tunes:

Rig "TAZ" @830W ATW:
6x Micron GPUs - 56.9MH/s @110W
Core clock: 1380
VDD: 769
Memory clock: 908
MVDD: 1264
MVDDCI: 762

Rig "COYOTE" @770W ATW:
6x Samsung GPUs - 57.2MH/s @100W
Core clock: 1398
VDD: 779
Memory clock: 912
MVDD: 1292
MVDDCI: 822

Rig "Cucumber" @735W ATW:
6x Samsung GPUs - 56.7MH/s @95W
Core clock: 1385
VDD: 766
Memory clock: 906
MVDD: 1286
MVDDCI: 772

All other - "Road_Runners" @760W ATW (same spec):
6x Samsung GPUs - 56.8MH/s @98W
Core clock: 1388
VDD: 773
Memory clock: 909
MVDD: 1314
MVDDCI: 790

Sensors averages:
GPU Edge: 47-51C
GPU Asic: 52-58C
GPU Mem: 71-84C (Cucumber/TAZ)

Credits to excellent work undertaken by @Mini_me, @AlleyCat and many talented others.

[Updated COYOTE]

 

[AlleyCat]

I'm sorry for late response.
I have RX 5700 XT Mech OC on all rigs now. GPUs in one rig are based on Micron memory, all the others - Samsung. I set them (all rigs) not to exceed 57MH/s as at the higher hash revenue ratio suffers and GPU temp goes above 82C. Naming convention from Looney Tunes:

Rig "TAZ" @830W ATW:
6x Micron GPUs - 56.9MH/s @110W
Core clock: 1380
VDD: 769
Memory clock: 908
MVDD: 1264
MVDDCI: 762

Rig "COYOTE" @770W ATW:
6x Samsung GPUs - 57.2MH/s @100W
Core clock: 1380
VDD: 769
Memory clock: 912
MVDD: 1264
MVDDCI: 762

Rig "Cucumber" @735W ATW:
6x Samsung GPUs - 56.7MH/s @95W
Core clock: 1385
VDD: 766
Memory clock: 906
MVDD: 1286
MVDDCI: 772

All other - "Road_Runners" @760W ATW (same spec):
6x Samsung GPUs - 56.8MH/s @98W
Core clock: 1388
VDD: 773
Memory clock: 909
MVDD: 1314
MVDDCI: 790

Sensors averages:
GPU Edge: 47-51C
GPU Asic: 52-58C
GPU Mem: 71-84C (Cucumber/TAZ)

Credits to excellent work undertaken by @Mini_me, @AlleyCat and many talented others.

Great work. I think the comparison between the hash rates is accurate. However, power values are too subjective.

Do you have a power meter to measure the power at the wall and compare the actual total power consumption between the rigs?

 

[AndSer]

Great work. I think the comparison between the hash rates is accurate. However, power values are too subjective.

Do you have a power meter to measure the power at the wall and compare the actual total power consumption between the rigs?

Very good question, I should have included the details in my report.

Each rig is individually connected to the power monitor (https://www.amazon.co.uk/gp/product/B07FZZ17ZY/) and this is where the ATW readings are taken from. I appreciate that these monitors are not lab grade calibrated, nonetheless reading values are fairly accurate.

From left - COYOTE, TAZ, Road_Runner#1 and #2

To the COYOTE's meter there is also a network switch connected which draws 5~8W.

 

[7pintwonder]

Hello, I have apple straps applied and am using MMPOS, but am only getting these settings. Attached are the bios mod and the settings I am using. Can someone help raise my MH and lower my wattage? Please let me know if you can help!

 

[7pintwonder]

I'm wondering if regular bios modding with 1500 down is better than Apple straps? I cant really tell. Seems the other method uses more power

 

[montecarlo]

I'm sorry for late response.
I have RX 5700 XT Mech OC on all rigs now. GPUs in one rig are based on Micron memory, all the others - Samsung. I set them (all rigs) not to exceed 57MH/s as at the higher hash revenue ratio suffers and GPU temp goes above 82C. Naming convention from Looney Tunes:

Rig "TAZ" @830W ATW:
6x Micron GPUs - 56.9MH/s @110W
Core clock: 1380
VDD: 769
Memory clock: 908
MVDD: 1264
MVDDCI: 762

Rig "COYOTE" @770W ATW:
6x Samsung GPUs - 57.2MH/s @100W
Core clock: 1398
VDD: 779
Memory clock: 912
MVDD: 1292
MVDDCI: 822

Rig "Cucumber" @735W ATW:
6x Samsung GPUs - 56.7MH/s @95W
Core clock: 1385
VDD: 766
Memory clock: 906
MVDD: 1286
MVDDCI: 772

All other - "Road_Runners" @760W ATW (same spec):
6x Samsung GPUs - 56.8MH/s @98W
Core clock: 1388
VDD: 773
Memory clock: 909
MVDD: 1314
MVDDCI: 790

Sensors averages:
GPU Edge: 47-51C
GPU Asic: 52-58C
GPU Mem: 71-84C (Cucumber/TAZ)

Credits to excellent work undertaken by @Mini_me, @AlleyCat and many talented others.

[Updated COYOTE]

Impressive result. Do you need to reapply the thermal pad on the mech oc? My 5600 XT Mech OC has bad stock thermal pads. Only 5 partially applied. Had to fix it with 6 fully covered thermal pad, which drops the mem temp down 20C

 

[AndSer]

Impressive result. Do you need to reapply the thermal pad on the mech oc? My 5600 XT Mech OC has bad stock thermal pads. Only 5 partially applied. Had to fix it with 6 fully covered thermal pad, which drops the mem temp down 20C

Maybe I was just lucky - out of all GPUs I sent back only one to the supplier. The said GPU very likely had some thermal issue right from the beginning as the fans went full throttle when powered up. Shy of 90C when idle. All the others - "tight and narrow".

 

[Mini_Me]

Once again I would like to thank all the contributors to this thread for the time and effort put in to testing. I do have my rig stable for the last couple of weeks. Now that I have a bit of time I would like to dial it in a little more. For those that are using Gigabyte 5700xt Gaming OC cards, can you post your voltage and memory settings? There is a lot of information in this thread, so I’m just looking for some final settings that you’ve come up with. I’m using the apple straps and MPT bios mod settings from this thread, and my hiveOS settings are below. I have a difference in wattage reported by 2 of my cards as they are the rev 2 of the same card, and thanks to another contributor to this thread it looks like that is just a software issue, actual wattage at the wall appears to be the same, so thanks for that as well!

I have now properly modified Pica Gigabyte RX 5700 XT OC rev.2 F3 vbios to bypass the rsa signature, this vbios will 100% work in CSM bios boot type however for UEFI bios boot type need testing,

Plrase give it a try and update whenever it is possible.

 

[Mini_Me]

I test the bios, don´t work, the GPU is bricked !

Now I properly modified the vbios that you sent me and it will work 100% on CSM bios boot type (OS on MBR partition) however the UEFI bios boot type (OS on GPT partition) needs testing,

Please give it a try and update whenever it is possible.

 

[AlleyCat]

Now I properly modified the vbios that you sent me and it will work 100% on CSM bios boot type (OS on MBR partition) however the UEFI bios boot type (OS on GPT partition) needs testing,

Please give it a try and update whenever it is possible.

The Power table looks the original. Can you confirm that the purpose of this bios is to handle the RSA signature?

 

[Mini_Me]

The Power table looks the original. Can you confirm that the purpose of this bios is to handle the RSA signature?

Yes as I already did this to an RX 5500 XT which I believe has better protection than this one.

 

[Mini_Me]

The Power table looks the original. Can you confirm that the purpose of this bios is to handle the RSA signature?

You can install HiveOS once as an UEFI system and once as a Legacy one and to change the boot type in the motherboard bios as well.