Searching stock bios for Sapphire PULSE RX 5700 XT BE

[HaelSturm42]

Hi all,

as I managed to loose the original stock rom of my card (please see title and reference links), am reaching out here in hope someone has it. Card is slightly different to normal RX 5700 XT as it uses different memory timings from what I can tell (it also lacks the dual bios switch and capability).

techpowerup specs
Sapphire product page

I already flashed following rom file from the database 224726 which lets me somehow partly interact (partly means it is visible to GPU-Z, reacts to API calls for OpenCL) with the card as long as it is only 2nd GPU, but leads to a black screen during windows and linux boot with a freeze of the complete system when I remove the other GPU (my old R7 used as primary).
Card is momentarily only used in that testing man cave rig, built with 650W PSU, MSI X58 mobo, Xeon L5645 CPU, 24 GB DDR3 RAM, 250GB SSD, Win10 (latest patch set), Ubuntu 24.04 LTS. 2nd card to have a screen is an old R7 XFX.

Thank you and best, Markus

 

[VuurVOS]

Hi all,

as I managed to loose the original stock rom of my card (please see title and reference links), am reaching out here in hope someone has it. Card is slightly different to normal RX 5700 XT BE as it uses different memory timings from what I can tell (it also lacks the dual bios switch and capability).

techpowerup specs
Sapphire product page

I already flashed following rom file from the database 224726 which lets me somehow partly interact with the card as Long as it is only 2nd GPU, but leads to a black screen during windows and linux boot with a freeze of the complete system when I remove the other GPU.
Card is momentarily only used in that testing man cave rig, built with MSI X58 mobo, Xeon L5645 CPU, 24 GB DDR3 RAM, 250GB SSD, Win10 (latest patch set), Ubuntu 24.04 LTS. 2nd card to have a screen is an old R7 XFX.

Thank you and best, Markus

You card uses board number E426. It is listed on PCB and on the label.

This bios will work likely since it shares the same board number (E426)

Sapphire RX 5700 XT VBIOS

8 GB GDDR6, 1400 MHz GPU, 1750 MHz Memory

www.techpowerup.com

 

[HaelSturm42]

@VuurVOS thank you for that sound advice - yes, this bios works. I needed to mend a 8 mbit file for amdvbflash to successfully finish this task.
These are my observations, in case someone finds this thread with similar challenges.

• amdvbflash 2.93 does not detect this card
• amdvbflash 3.04e does
• flashing with force parameters still yields a failure
  • I believe this is caused by the fact, the card is equipped with GD25Q80 flash (2x512k or 8 mbit size) and the bios rom files is only 512k
  • Generally a bios switch would switch a performance or silent bios bank, the BE is not equipped with this switch
  • attempts to read the flash yield a 1024 kb file - obviously first 512k contain the bios, the rest contains FF
  • hence it seems to me this size mismatch leads to flash failure
• flashing did succeed if a 1024k (or 8 mbit) file is used though (attached as zip)
  • I mended such file in a hex editor: 4 mbit with suggested bios followed by 524800 bytes FF
• results look ok, please see attached GPU-Z screenshots
• still the card causes a black screen during boot attempt on X58 Westmere Xeon mobo with latest AMD Adrenaline drivers BUT it does flawlessly work on another board with newer chipset and CPU
• coming from cMP 4,1 and 5,1 Westmere XEON systems I can only assume new AMD drivers make in fact use of AVX2 instructions for Win10, too, not only newer Mac OS as of Ventura. Which this old platform does not support leading to a black screen as soon the driver loads. During Bios POST there is a UI displayed until Win10 loader takes over and loads drivers.

 

[eidairaman1]

@VuurVOS thank you for that sound advice - yes, this bios works. I needed to mend a 8 mbit file for amdvbflash to successfully finish this task.
These are my observations, in case someone finds this thread with similar challenges.

• amdvbflash 2.93 does not detect this card
• amdvbflash 3.04e does
• flashing with force parameters still yields a failure
  • I believe this is caused by the fact, the card is equipped with GD25Q80 flash (2x512k or 8 mbit size) and the bios rom files is only 512k
  • Generally a bios switch would switch a performance or silent bios bank, the BE is not equipped with this switch
  • attempts to read the flash yield a 1024 kb file - obviously first 512k contain the bios, the rest contains FF
  • hence it seems to me this size mismatch leads to flash failure
• flashing did succeed if a 1024k (or 8 mbit) file is used though (attached as zip)
  • I mended such file in a hex editor: 4 mbit with suggested bios followed by 524800 bytes FF
• results look ok, please see attached GPU-Z screenshots
• still the card causes a black screen during boot attempt on X58 Westmere Xeon mobo with latest AMD Adrenaline drivers BUT it does flawlessly work on another board with newer chipset and CPU
• coming from cMP 4,1 and 5,1 Westmere XEON systems I can only assume new AMD drivers make in fact use of AVX2 instructions for Win10, too, not only newer Mac OS as of Ventura. Which this old platform does not support leading to a black screen as soon the driver loads. During Bios POST there is a UI displayed until Win10 loader takes over and loads drivers.

That x58 is up to latest bios too though?

 

[HaelSturm42]

Yeah, it is. It is an old MSI X58 PRO-E ATX with bios X58 Pro SLI (MS-7522) V8.15 BIOS Release
LGA1366 Xeon CPUs (here L5645) don't support AVX and neither AVX2 unfortunately. It might as well be an incompatibility with that particular CPU / mainboard / GPU.

I might try some linux dkms AMD drivers (which should build without unsupported instructions) out of curiosity, but this thread was about bios flashing in the first place

 

[eidairaman1]

Yeah, it is. It is an old MSI X58 PRO-E ATX with bios X58 Pro SLI (MS-7522) V8.15 BIOS Release
LGA1366 Xeon CPUs (here L5645) don't support AVX and neither AVX2 unfortunately. It might as well be an incompatibility with that particular CPU / mainboard / GPU.

I might try some linux dkms AMD drivers (which should build without unsupported instructions) out of curiosity, but this thread was about bios flashing in the first place

Yeah I was just making sure bases are covered as I did write the bios flash guide for ATi cards, i wrote it because i got tired of writing them down in every thread lol. Since it works else where Safe to say chalk it up to that.

 

[HaelSturm42]

Yeah I was just making sure bases are covered as I did write the bios flash guide for ATi cards, i wrote it because i got tired of writing them down in every thread lol. Since it works else where Safe to say chalk it up to that.

Got ya and fully agree on that