Polaris bios editor ошибка

Send us the bios. H5GC8H24AJ was pretty common in 2017, so not sure what the problem is.

See attachment.See problem in attachment.See useful linkhttps://miningclub.info/threads/pamjat-h5gc8h24aj.56428/  21.01.2021, 23:06, «Indeed Miners» <notifications@github.com>: Send us the bios. H5GC8H24AJ was pretty common in 2017, so not sure what the problem is.—You are receiving this because you authored the thread.Reply to this email directly, view it on GitHub, or unsubscribe.

There is no attachment, you probably need to upload directly on Github not as email attachment. The thread is also 2 years old and I don’t speak russian (Google Translate is not speaking very good russian too)
My bad I’ve confused H5GC8H24AJ for H5GC4H24AJ, but still both of them are supported in PBE. H5GC8H24AJ since 1.7.2.

https://yadi.sk/i/OBQ66JMxQA6mAw — problemhttps://yadi.sk/d/dFJemjZf1d3vvA — bios Dont read link, you can only test timings for this memory type.  22.01.2021, 14:46, «Indeed Miners» <notifications@github.com>: There is no attachment, you probably need to upload directly on Github not as email attachment. The thread is also 2 years old and I don’t speak russian (Google Translate is not speaking very good russian too)—You are receiving this because you authored the thread.Reply to this email directly, view it on GitHub, or unsubscribe.

Thanks for the bios. I can reproduce the error. It works just fine, it applies correct timings. We will fix the error alert in next release. PBE PRO is completly fine with that, there is no error at all.

Polaris Bios Editing RX590/RX580/570/560/RX480/470/460
Last update: 12/03/2021

New Drivers Allow modded Bios on RX580/590, RX480 needs «Atikmdag — Patcher» to fix the Code 43 error/Drivers not loading

Atikmdag-Patcher — Bios Enabler — Code 43 Fix

RX480 > RX580 Conversion

PBE — Polaris Bios Editor — Versions + Infos

Timings Decoding

GDDR5 Memory Timings Info

####SEQ_RAS_TIMING####
TRCDW = 13   “Number of cycles from active to write”
TRCDWA = 13  “Number of cycles from active to write with auto-precharge. Same as TRCDW” Good BW uplift???
TRCDR = 24   “Number of cycles from active to read”
TRCDRA = 22  “Number of cycles from active to read with auto-precharge. Same as TRCDR”
TRRD = 5     “Number of cycles from active bank a to active bank”
TRC = 65     “Number of cycles from active to active/auto refresh”

####SEQ_CAS_TIMING####
TNOPW = 0   “Extra cycle(s) between successive write bursts. For debugging purpose only”
TNOPR = 0   “Extra cycle(s) between successive read bursts. For debugging purpose only”
TR2W = 28   “Read to write turn around time”
TCCDL = 3   “Cycles between column commands between banks in the same bank group?”
TCCDS = 5   “Cycles between column commands between banks in different bank groups?”
TW2R = 14   “Write to read turn around time”
TCL = 20    “CAS to data return latency”

####SEQ_MISC_TIMING####
TRP_WRA = 45   “From write with auto-precharge to active”  Good BW uplift???
TRP_RDA = 22   “From read with auto-precharge to active”
TRP = 19       “Precharge command period”
TRFC = 151     “Auto-refresh command period”

####SEQ_MISC_TIMING2####
PA2RDATA = 0   “Read Preamble”
PA2WDATA = 0   “Write Preamble”
TFAW = 0       “Four Bank Active Window”
TCRCRL = 3   EDC READ Latency=CL + CRCRL “During read ,the EDC bundle is returned with-soon after the response data based on TCRCRL”
TCRCWL = 7   EDC WRITE Latency=WL + CRCWL "EDC/CRC checksum Related, similar to TCRCRL"
T32AW = 0      “Thirty Two Bank Active Window”
TWDATATR = 0

####ARB_DRAM_TIMING####
ACTRD = 30
ACTWR = 18
RASMACTRD = 58
RASMACTWR = 70

####ARB_DRAM_TIMING2####
RAS2RAS = 219
RP = 53
WRPLUSRP = 64
BUS_TURN = 25

How to Read/Mod Memory HEX Values

UberMix and Performance Timings

777000000000000022CC1C00EF615C44F0590F15300D070A0060070013051420FA8900A0030000001212303FB2354019

777000000000000022CC1C00AD615C42F0590F15300D9708006007000B031420FA8900A00300000011112F3FBB354019

777000000000000022CC1C00AD615C41C0590E152ECC8608006007000B031420FA8900A00300000010122F3FBA354019

777000000000000022CC1C00AD615B41C0570E152DCB7508006007000B031420FA8900A00300000010123A46B9354019

777000000000000022CC1C00AD615B41C0570E152DCB7409006007000B031420FA8900A00300000010123A46DB354019

555000000000000022CC1C00AD615B41C0570E152DCB74090068C7000B031420FA8900A0030000001B11333DC0303A17

216GB/s Stock (Needs update)
14200 FireStrike Stock
THE WITCHER 3(Ultra-No HW-AA On) 53fps 1080P Stock
THE WITCHER 3(Ultra-No HW-AA On) 22fps 4K Stock

229GB/s MOD +6% (Needs update)
14500 FireStrike MOD +2.1%
THE WITCHER 3(Ultra-No HW-AA On) 56fps 1080P MOD +5.6%!
THE WITCHER 3(Ultra-No HW-AA On) 23fps 4K MOD +4.5%

EDC Errors monitoring

Polaris Memory Overclock Scaling

Polaris 10 Voltage Scaling

Good Reads/Articles and Infos

Random Info

Downloads

Make backup of original bios on video card. Using modded bios will void your warranty. I accept no responsibility for damage from using this information. All efforts are being made to double check information but there maybe errors.

See less

See more

Reactions:
1

Hi!

I am trying to edit bios for my gpus so I never need to setup overclocking and undervolting for cards separately. So I have tried to edit bios for 2 of my cards
SAPPHIRE NITRO+ Radeon™ RX 570 4GD5
SAPPHIRE PULSE ITX Radeon™ RX 570 4GD5

There’s everything fine with overclocking. But I am failing to do undervolting correctly. Basically I can not undervolt core and memory voltage. It does not matter what voltage I am putting in bios editors I am always having high power consumption (110-130) and no change in VDDC. Power consumption gets better if I lower TDP and power limit — that gives some results, but still core settings does not give any results. But when undervolted in MSI AB it works perfectly like 90w in full load on Claymores dual miner eth+dcr 27mhs+300mhs

I am always modifying my original bios. Using only one type of bios editor on each try. I have tested all those bios editors listed here:
PolarisBiosEditor v1.5 Epsylon3
PolarisBiosEditor v1.4
SRBPolaris V3 by d0kt0r
PBE 1.5

Drivers on Windows does not make any change
I have made my own which are working fine on several rigs by this guide
http://1stminingrig.com/run-6-rx-570-rx-580-gpus-windows-10/
I have tested with 1.7.7.2 and 1.7.7.1 with atikmdag-patcher.

Of course I did DDU every time switching drivers, My Windows is up to date. And same problem is on completely different PC. Also I do power cycle GPU after flashing BIOS.

Flashing with atiflash 274 and 277

Also changed values for FAN does not seem to work for me but that’s not my concern right now.

This is how my undervolted bios looks like in real situation.
https://drive.google.com/file/d/0B1u6LMtF_uAVdG1kbzdmNFFYVzg/view?usp=sharing

And this is what i did on bios editor
https://drive.google.com/file/d/0B1u6LMtF_uAVZ3VGanlIVVRDUTg/view?usp=sharing

But that’s only one of the variations I did. No matter what I put in bios file with which BIOS editor it is always possible to set in AB core -96 and it still will be stable, which means, no change on VDDC has been done. Also I hev set very low values in mem und core like 850 or so — still — no change, cards working stable, which I doubt they would with such low voltage.

Sometimes i get glicth that it is working finaly, GPUZ and AB shows voltage 900mv and power consumption is 85-90w but reflashing same bios or restarting pc does to give those results anymore.

Could you please point me where I could been mistaken? What did I miss? Or are those cards locked for voltage control?

@deffdsp

I have the nitro card and I did a similar voltage edit in the jaschaknack fork of PBE (maybe the same as what you call PBE 1.5? I dunno) but it works for me. The only difference I see though is that you also edited the frequencies of the higher DPM states as well as the voltages, but that shouldn’t matter especially since you’re not running higher than 1168 core clock anyway.

I noticed I can’t set voltage to anything lower than 950 though and have it work….would probably have to go into editing VOI table with hex and recalculating table lengths, etc. That doesn’t seem like it’s be necessary though because less than that is unstable anyway. Maybe 925 is just being ignored because it’s too low? What if you try 950 or 1000?

If you find out how to set the fan curve in BIOS though please tell me because I would love to know that also

Quote:

Originally Posted by Vento041

…

• Example: the 1750 strap is active from 1750 to 1999, at 2000 the 2000 strap will be selected

…

Again, you may want to double check this statement. I’m quite sure the 1750 strap is active from 1626 — 1750, and the 2000 strap is active from 1751 — 2000 (assuming you don’t have an 1875 strap in between). This (the way I’ve described) is how it works on Hawaii, and also for at least one of my 580 cards. It isn’t too stable above 2000 so I’ve run it at 1980-1990, and it appears to be using the timings I’ve set in the 2000 strap, not the stock 1750 strap.

2

Quote:

Originally Posted by PPBottle

You need to extract BIOS’es with the Atiwinflash itself (save option), that way you always get the full 512kb bios.

GPU z only extracts the bits with stored data. Some polaris bioses have 256kb blank space and 256kb actually written, this is why you get that error on Polaris Bios editor.

Wait, how are you extracting the full bios via the save option from Atiwinflash? To date I had never run into the bios being smaller than 512 and now Im seeing it on every card I get my hands on. Had no problems with my 570, 480, and a 470. Now a new 580 and a new 470 are both giving me a hard time. Even opening one of them gives misread info no matter what profile for vram is selected in hte drop down. Ive been given two different 512bit bios from friends for the 580. Both are claiming they are less than 512bits. The one 512bit I got for the 470 wont flash do to subsystem id mismatch. With that I did nothing more than copy my 1500mhz timings to the other frequencies. Im just totally lost as to why every step gives me a problem when I had no problems before

Has anyone run across cards that have 2 different memory manufacturers?? I recently acquired 6 new Dell AMD RX480 reference cards, but PBE 1.5 read the stock rom with strange numbers and dropdown menu says «2 Brands». SRBPolaris v3 warns unknown model when loading the rom and shows Samsung and Hynix. See screenshots below:

PBE 1.5:

SRBPolaris:

Stock ROM:
https://www.dropbox.com/s/6qnkb6dwzufv7gd/stock0.rom?dl=0

5

Quote:

Originally Posted by Vento041

@DMatthewStewart

• There are no real «512kb» bios, they are just 256kb bios + another 256kb of empty space.
• The misread values can be caused by outdated PolarisBiosEditor version, since some of them have «hardcoded» values/offsets to read the bios, so if something is different the app trip and show you the wrong bios section.
• The 1500MHz strap copy to higher strap is a lazy (not) solution on how mod memory timings read my post above and if you really have time read my R_Timings project.

Thank you sir. Turns out, for some reason unknown, ATI Win Flash was not letting me save the BIOS of any card. The «Save» button was just blank. And for some other unknown reason, it now comes up and is useable. So I’ve been able to extract the 512byte ROM. But youre right, the 256kb ones are no different. Theyre just 50% empty

And I agree that copying the straps is lazy sometimes. But for my application its strictly to run a mining rig. Works well for the other cards. Actually, for most it turns out the best results Ive seen yet. In a few instances other straps have produced a marginal gain over the copying of the 1500 straps but its not the norm. Matter of fact, the last vbios I modded myself was a touch faster on memory. It worked for about a week and began to fail. Went back to using my 1500 straps at 1750 and 2000 and its stable for my application. We’re talking about zero 3d output, memory intensive calculations only

But I will take some time and read your R_Timings Project, and I thank you for that and your other suggestions.

Quote:

Originally Posted by Leons

Quote:

Originally Posted by bardacuda

Ok perhaps I am mistaken. I will re-test when I get that particular card back in my main rig for further tuning.

Imho you are right

I still haven’t tested this so don’t take what I said at face value. I’m still pretty sure it works the way I explained but I’m a little less sure now with doktor also chiming in. He and Vento both know their stuff, so basically just try it out for yourself. In the end it usually won’t come into play unless you are straddling the 2000 MHz (or 1750 MHz, if I am right) border. Most people are running > 2000 MHz I imagine.

EDIT: Actually I am gonna test this right now. I am in the middle of testing experimental timings so why not. I will edit only the 2000 MHz strap, leaving the 1750 strap at stock, and flash. Then I will reboot and test at 1749 MHz, reboot again and test at 1751, reboot, 1999, reboot, 2001.

If I am right I should see a big increase at 1751 vs. 1749 and no difference at 2001 vs. 1999. If they are right then I should see a big increase at 2001 vs 1999, and no difference a decrease at 1751 vs. 1749 (due to using the looser 1750 strap at 1751 and the tighter 1625 strap at 1749). More to come…

EDIT2:

Results:

• 1749: 20.5MH/s

  

• 1751: 21.0MH/s (Not as much increase as I thought, but my timings are geared for 2150, and it’s not a decrease)

  

• 1999: 25.6MH/s (Clearly using the experimental timings here already. No way I would get this with stock timings.)

  

• 2001: 25.8MH/s (It increased 0.2MH/s which I actually can’t explain, except that I have noticed some variance between boot cycles before. This means the increase between 1749 and 1751 is also kind of meaningless though.)

  

And I’ll throw in my 2150 results too since I am testing it atm anyway.

• 2150: 28.0MH/s (Wish it were higher

  )

  

So anyway it may not be entirely definitive…but for me the results at 1999 are all I need to know that it is using the 2000 MHz strap. I already did stock testing and my hash rate was way lower, even with a higher core frequency and on an earlier epoch (hash rate goes down with each epoch…at least until AMD fix their drivers)

http://www.overclock.net/t/1604567/polaris-bios-editing-rx5xx-rx4xx/#post_26231584

TL;DR: Don’t bother editing your 1750 strap (unless for some reason you are running your memory <= 1750). This way you have a better chance to avoid soft bricks.

6

Quote:

Originally Posted by DMatthewStewart

I know. I should. But I just get too lazy. Every once in a while I want a gui. Plus, I hardly remember Win commands. My brain is filled with other commands from better os’s

write:
atiflash.exe -p 0 bios_to_write.rom

if you do some 4xx<->5xx conversion (or you triggered any ID mismatch error on purpose)
add «-f» at the end

save:
atiflash.exe -s 0 where_to_save.rom

info from running card:
atiflash.exe -ai 0

info from bios file:
atiflash.exe -biosfileinfo bios.rom

Quote:

Originally Posted by bardacuda

I still haven’t tested this so don’t take what I said at face value. I’m still pretty sure it works the way I explained but I’m a little less sure now with doktor also chiming in. He and Vento both know their stuff, so basically just try it out for yourself. In the end it usually won’t come into play unless you are straddling the 2000 MHz (or 1750 MHz, if I am right) border. Most people are running > 2000 MHz I imagine.

EDIT: Actually I am gonna test this right now. I am in the middle of testing experimental timings so why not. I will edit only the 2000 MHz strap, leaving the 1750 strap at stock, and flash. Then I will reboot and test at 1749 MHz, reboot again and test at 1751, reboot, 1999, reboot, 2001.

If I am right I should see a big increase at 1751 vs. 1749 and no difference at 2001 vs. 1999. If they are right then I should see a big increase at 2001 vs 1999, and no difference a decrease at 1751 vs. 1749 (due to using the looser 1750 strap at 1751 and the tighter 1625 strap at 1749). More to come…

EDIT2:

Results:

• 1749: 20.5MH/s

  

• 1751: 21.0MH/s (Not as much increase as I thought, but my timings are geared for 2150, and it’s not a decrease)

  

• 1999: 25.6MH/s (Clearly using the experimental timings here already. No way I would get this with stock timings.)

  

• 2001: 25.8MH/s (It increased 0.2MH/s which I actually can’t explain, except that I have noticed some variance between boot cycles before. This means the increase between 1749 and 1751 is also kind of meaningless though.)

  

And I’ll throw in my 2150 results too since I am testing it atm anyway.

• 2150: 28.0MH/s (Wish it were higher

  )

  

So anyway it may not be entirely definitive…but for me the results at 1999 are all I need to know that it is using the 2000 MHz strap. I already did stock testing and my hash rate was way lower, even with a higher core frequency and on an earlier epoch (hash rate goes down with each epoch…at least until AMD fix their drivers)

http://www.overclock.net/t/1604567/polaris-bios-editing-rx5xx-rx4xx/#post_26231584

TL;DR: Don’t bother editing your 1750 strap (unless for some reason you are running your memory <= 1750). This way you have a better chance to avoid soft bricks.

Nice that you tested this, now if you have the time, do the same test for 1999 but return stock 2000mhs timings, leave everything else same.
Im curious what will happen, as i never tested this so thoroughly

1

Why not?

EDIT:

Result:

• 1999 Stock Timings: 21.1MH/s (Slightly less than my stock BIOS testing as expected which was done at 1366/2000 on Epoch 137)

  

EDIT2: Btw this testing was done on the MSI Armor 8GB card which I think I failed to mention. Oh btw @doktor83 I made an error and put >= 1750 instead of <= 1750 but I’ve fixed it. Can you fix it in your quote of my post too? The whole point of my testing was to clear up any confusion and that could just add more

Thanks!

2

Quote:

Originally Posted by bardacuda

Why not?

EDIT:

Result:

• 1999 Stock Timings: 21.1MH/s (Slightly less than my stock BIOS testing as expected which was done at 1366/2000 on Epoch 137)

  

EDIT2: Btw this testing was done on the MSI Armor 8GB card which I think I failed to mention. Oh btw @doktor83 I made an error and put >= 1750 instead of <= 1750 but I’ve fixed it. Can you fix it in your quote of my post too? The whole point of my testing was to clear up any confusion and that could just add more

Edited post

This test looks interesting. I never tried -1MHZ to be honest, i tried -10Mhz (1990), on a sapphire nitro+ 470 samsung mem card, and then it wasn’t using the 2000 strap, cause i got the extra boost only on >= 2000.
So that is why this is interesting.

@Vento041 , what are your thoughts about this?

1