Question GIGABYTE 5700 XT Bios mod fails

AlleyCat

Veteran
Mitglied seit
Okt 9, 2020
Beiträge
290
Bewertungspunkte
17
Punkte
17
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
 
Zuletzt bearbeitet :
Yes, it looks like it is stable.
Let us revert back to the first power table of before as below with a little increase in the SoC TDC, this is for Sapphire card with Samsung memory only and this is not temporary:

[Updated]
This is the final power table for any RX 5700 card in the kindly attached link below,

 
Zuletzt bearbeitet :
And regarding the vram timings,
We are going to use Apple straps and to copy the 2000 MHz Apple Strap to the 1800 MHz strap while preserving their tREF values, this way we are now starting to tighten the vram timings bit by bit.
At 960 the card crashes. At 955 it looks like it is working, but with an Invalid shares.
Changing the card speed to lower crashes the card.
 
Zuletzt bearbeitet :
At 960 the card crashes. At 955 it looks like it is working, but with an Invalid shares.
Changing the card speed to lower crashes the card.
I believe it is for the best to apply the Apple vram timings without any modification and we increase the VDDCI and MVDD voltages until we reach stable memory clock at 925 MHz or 930 MHz.
 
I believe it is for the best to apply the Apple vram timings without any modification and we increase the VDDCI and MVDD voltages until we reach stable memory clock at 925 MHz or 930 MHz.
OK. I am reverting to Apple's straps, and the below power table.
I suggest to make one change. To increase the GFX Maximum Clock to 1450, because my system is stable at 1420MHz core clock. Perhaps the final settings after power optimization will be below 1400MHz, but perhaps having the headroom could help with tuning.

Under Features tab enable all the overdive features,
- Under Overdrive Limits tab, set:
GFX Maximum Clock 1400 MHz
Memory Maximum Clock 1000 MHz
Under Power and Voltages, set:
Maximum voltage GFX 850 mv
Maximum voltage SoC 1200 mv
Minimum voltage GFX 750 mv
Minimum voltage SoC 750 mv
Power Limit GPU 150 W
TDC Limit GFX 136 A
TDC limit SoC 14 A
- Under Frequency tap, set:
GFX maximum 1400 MHz
Memory DPM 3 - 960 MHz
- Under Fan tab, set:
Fan Acoustic Limit 1380 RPM
Fan throttling 2900 RPM

Like Quote answer
report
 
I believe it is for the best to apply the Apple vram timings without any modification and we increase the VDDCI and MVDD voltages until we reach stable memory clock at 925 MHz or 930 MHz.
I am getting Invalid Shares at memory clock 925, and at memory clock 930 the card crashes.
In both cases my VDDCI = 900 and MVDD = 1390
 
I am getting Invalid Shares at memory clock 925, and at memory clock 930 the card crashes.
In both cases my VDDCI = 900 and MVDD = 1390
Let us first apply the Apple vram timings without any modification with stock voltages and after that we increase the voltages until we reach stable 925 MHz or 930 MHz.
 
Let us first apply the Apple vram timings without any modification with stock voltages and after that we increase the voltages until we reach stable 925 MHz or 930 MHz.
I.am not sure I understand.
What is the stock voltage? are you referring to VDD? VDDCI? MVDD? At what values?

I know that MVDD = 1.35V (1350mv)
What are the other stock values?
 
Stock voltages are the default ones like the MVDD at 1350 mv, VDDCI at 750 mv minimum, VDD at 750 mv minimum.
OK. So what should be the test?
I start at memory clock 900 and start increasing until the system returns invalid shares errors.
How do you you suggest I increase the MVDD and VDDCI? In what increments, and what are the max values?
 
OK. So what should be the test?
I start at memory clock 900 and start increasing until the system returns invalid shares errors.
How do you you suggest I increase the MVDD and VDDCI? In what increments, and what are the max values?
Let us finalize this card by making a moded bios with the last power table for Sapphire with Samsung memory and Applying the Apple vram timings after flashing we increase the MVDD and VDDCI until we reach stable 925 MHz or 930 MHz.
We would target higher like 955 MHz however I am afraid that MVDD more than 1370 mv will degrade the memory longevity and performance.
 
OK. So what should be the test?
I start at memory clock 900 and start increasing until the system returns invalid shares errors.
How do you you suggest I increase the MVDD and VDDCI? In what increments, and what are the max values?
I think we need to stabilize the card for production. It is hard for me to consider higher speeds from the card with the Samsung memory.
However, I have an interesting discovery. The card's efficiency is best at memory clock 910MHz, with Core 750mv, and VDDCI 750mv. The power went down to 97 - 97 W.
Please check these numbers. Would you agree that getting 57MH/s at 97W (reported by HiveOS) is efficient? I also have to consider stability and multiple cards variations.

I think Apple's straps work steady and give stable increase of hash speed.

How about we shift the focus to reducing the power for all cards?
 

Anhänge

  • Igors Test Report.png
    Igors Test Report.png
    171,1 KB · Aufrufe : 186
I think we need to stabilize the card for production. It is hard for me to consider higher speeds from the card with the Samsung memory.
However, I have an interesting discovery. The card's efficiency is best at memory clock 910MHz, with Core 750mv, and VDDCI 750mv. The power went down to 97-97 W.
Please check these numbers. Would you agree that getting 57MH / s at 97W (reported by HiveOS) is efficient? I also have to consider stability and multiple cards variations.

I think Apple's straps work steady and give stable increase of hash speed.

How about we shift the focus to reducing the power for all cards?
I totally agree with you, we tried our best however the memory controller is not optimized and locked intentionally by AMD.
 
I think we need to stabilize the card for production. It is hard for me to consider higher speeds from the card with the Samsung memory.
However, I have an interesting discovery. The card's efficiency is best at memory clock 910MHz, with Core 750mv, and VDDCI 750mv. The power went down to 97 - 97 W.
Please check these numbers. Would you agree that getting 57MH/s at 97W (reported by HiveOS) is efficient? I also have to consider stability and multiple cards variations.

I think Apple's straps work steady and give stable increase of hash speed.

How about we shift the focus to reducing the power for all cards?
The 57 MH/s at 97 watt is a brilliant choice.
 
Oben Unten