@CabalNarrow 1750 is what I'll optimise for with a focus on ensuring stability without any impact worth noting on performance while also trying to maximise whatever OC capability those ICs have and keeping them stable. Its going to be best to treat the memory like Hynix/Micron with timings but with the characteristics of OCing Samsung. In short, this is going to be tricky balancing everything patience and testing is key here sticking to the straps in the modded firmware images is going to make this a lot simpler to work out and get a feel for these ICs to isolate exactly whats upsetting them chopping and changing completely different memory straps is only going to exponentially complicate the process. Wipe out the driver with DDU and test these timings at 1750 for at least an hour and assuming there are no errors bump up 50MHz at a time and test until you start getting EDCs.

@ket
Thank you as always for your time. Had been able to test it only today and it feels stable at 1900 (1hour FFBench followed by 1 Firestrike and 1 Time Spy) with no errors.
Will test it a bit more tomorrow if I'm able to due to work and get it back to you!

@CabalNarrow I think you probably have an oddball card I have one just like yours where it even threw EDCs at stock timings and frequencies. Ironically I cleared a lot up by using actual good values and not the terrible stock timings, didn't resolve the EDCs fully until after a couple days of testing to work out what the problem was. I'm not sure if in both the oddballs cases if the issue is a slightly iffy memory IC or memory controller but I did work out the problem was with sequential CAS timing settings luckily those settings are pretty forgiving you need to adjust them a long way before there is any kind of noticeable performance drop. With the changes I needed to make to my oddball card its maybe at worst about 1FPS slower than a "normal" RX470 at the same clocks.

@ket
I think I tryied stock timings too and they were giving errors as well. Even with the timing now you gave me there's no much difference in in-game FPS (Red Dead Redemption 2 and Warzone), guess I will try to tight it the most I can and then apply a bit of Core overclock.

@CabalNarrow thats the thing, its the bad stock timings that cause EDCs in some cases. Particularly TCCDL, for whatever warped reason AMD set it to 3 and from all whitepapers I've been able to get my hands on it should be set to at least 4 and in particularly troublesome cases use 6.

Excellent guide really loved reading it. I tried some timings and saw some nice increments but now I dont know what happened but if I change memory clock it doesn't reflect back in OCL I think its related to drivers.
Like it gives me 185Gbytes/s at 1750mhz(rx570 4gb samsung memory) but if i crank it to 2000mhz it stays at 185Gbytes/s and if i make it 2100mhz its remains the same

Zitat von wildcard im Beitrag #156

---

Excellent guide really loved reading it. I tried some timings and saw some nice increments but now I dont know what happened but if I change memory clock it doesn't reflect back in OCL I think its related to drivers.
Like it gives me 185Gbytes/s at 1750mhz(rx570 4gb samsung memory) but if i crank it to 2000mhz it stays at 185Gbytes/s and if i make it 2100mhz its remains the same

---

That fixed it Thank you. Really had my head scratching.
I have a question, I have a Gigabyte RX570 with Samsung K4G41325FE-HC25 which are rated for 8gbps but my card runs at 1750Mhz instead of 2000Mhz and when I try to run them at 2000Mhz they run fine for sometime and then whole computer crashes. What exactly is holding it back thats confusing me so much.
I have attached the bios that I am currently running. I have only modified 1900 Strap 1750 and 2000 are original it is stable at 1900Mhz memory clock please take a look

@wildcard The imc of the rx570 chip holding you back.

Changing the VDDCI will help but how do I do the card has uP9505P. Is the procedure same as in Guide?

Zitat von wildcard im Beitrag #158

---

That fixed it Thank you. Really had my head scratching.
I have a question, I have a Gigabyte RX570 with Samsung K4G41325FE-HC25 which are rated for 8gbps but my card runs at 1750Mhz instead of 2000Mhz and when I try to run them at 2000Mhz they run fine for sometime and then whole computer crashes. What exactly is holding it back thats confusing me so much.
I have attached the bios that I am currently running. I have only modified 1900 Strap 1750 and 2000 are original it is stable at 1900Mhz memory clock please take a look

---

Hi! I’m working on app for rebuilding modules in vbios. And early method working fine. But I stuck with vbios signature verification. I compare all modules of four vbioses and find out that PowerPlayInfo includes some different two bytes 3125, 3025, 4626, 6326... etc... so possibly they are signatures.

[img][/img]


Then i tried to understand Vbios header, and that is what i made:

Zitat

---

55 AA FILE MASK
7F E9 FILE SIZE

A1 02 OFFSET TO “MP” FILE INSIDE VGA FIRMWARE ??????????????????????????

00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

44 02 OFFSET TO PCIR (0x0244)

00 00 00 00

49 42 4D IBM (ASCII)

04 CRC-8 FileChecksum (Based on Filesize E97F)

3A 88 OFFSET TO PFR IMAGE ??????????????????????????????

00 00 00 00 00 00 00 00 00 00 00

04 20

37 36 31 32 39 35 35 32 30 761295520

00 00 00 00 00 00

99 02 OFFSET TO VGA FIRMWARE (0x0299)

00 00 00 00 00 00

20 02 OFFSET TO ATOM SIGNATURE (0x0220) (ATOM)

00 00 00 00 00 00

31 32 2F 30 32 2F 31 36 2C 31 30 3A 31 34 3A 30 30 12/02/16,10:14:00 - RELEASE DATE (loc 0x0050) (ASCII)

00 00 00 E9 A7 03 00 E9 B1 03 00 00 00 F4 00 00 15 00 00 00 D0 01 00 AA 01 21 E1 02 80 7E 00 22 0F 44 02 12 00 00 00 00 00 00 3C 40 0E 02 07 3C 01 1A 00 04 00 00 00 EE A0 FF 06 00 08 30 40 0E 01 00 00 00 00 00 00 14 03 00 00 00 00 00 00 BE 7E 11 00 B9 07 1A D6 50 2C 00 00 00 00 00 00 00 00 00 00 0C 40 41 43 00 00 00 00 10 00 00 00 42 00 00 00 80 7E 06 00 20 00 20 00 12 00 0E 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 42 52 30 31 30 39 30 2E 30 30 31 00 46 47 4C 20 42 41 46 46 49 4E 4D 00 50 43 49 5F 45 58 50 52 45 53 53 00 47 44 44 52 35 00 0D 0A 48

50 20 56 61 75 67 68 6E 20 47 31 2D 35 30 20 47 44 44 52 35 HP Vaughn G1-50 GDDR5 (Product Name) (ASCII)

20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20

0D 0A 00 0D 0A 20 0D 0A 00

28 43 29 20 31 39 38 38 2D 32 30 31 30 2C 20 41 64 76 61 6E 63 65 64 20 4D 69 63 72 6F 20 44 65 76 69 63 65 73 2C 20 49 6E 63 2E (C) 1988-2010, Advanced Micro Devices, Inc. (ASCII)

00

41 54 4F 4D 42 49 4F 53 42 4B 2D 41 4D 44 ATOMBIOSBK-AMD (ASCII)

20

56 45 52 VER (ASCII)
30 31 35 2E 30 35 30 2E 30 30 30 2E 30 30 30 2E 30 30 31 30 39 30 015.050.000.000.001090 (BIOS VERSION) (ASCII)

00 42 52 30 31 30 39 30 2E 30 30 31 20 00 31 33 34 38 38 34 37 20 00 33 36 38 31 32 34 20 20 00 20 20 20 20 20 20 20 20 00 48 50 5F 42 41 46 46 49 4E 4D 5F 47 4C 50 52 4F 5F 56 41 55 47 48 4E 5F 47 44 44 52 35 5F 4D 58 4D 5C 63 6F 6E 66 69 67 2E 68 00 00 00 90



ATOM ($...ATOM.АЕКiк….<ХЂD†‹,Њ .)
24 00 FileSize (0024)
01 01 Version (0101)
41 54 4F 4D ATOM (ASCII) (Atom Magic) uaFirmWareSignature[4]
00 C0 OFFSET TO BIOS Runtime Segment Address (0x00C0)
C5 03 OFFSET TO ProtectedModeInfo (0x03C5) (PM IMAGE)
CA 01 OFFSET TO ConfigFileName (0x01CA) BR01090.001 (ASCII)
69 02 OFFSET TO CRC_Block (0x0269) (HEX = B7 28 39 10)
1E 01 OFFSET TO BiosBootupMessage (0x011E)
EA 03 OFFSET TO (0x03EA) Int 10
00 00 PCI Bus Device Init Code
00 00 IO Base Address
3C 10 SUBVENDOR ID (HP)
D5 80 SSID (80D5)
44 02 OFFSET TO PCIR (0x0244)
86 8B OFFSET TO COMMAND TABLE (0x8B86)
2C 8C OFFSET TO DATA TABLE (0x8C2C)
A0 Extended Function Code
00 Reserved



PCI CONFIG

50 43 49 52 PCIR (ASCII)
02 10 VENDOR ID
E8 67 DEVICE ID
00 00
18 00
00 00
00 03
7F 00
32 0F
00 00
00 00
41 4D
44 20
41 54 4F 4D 42 49 4F 53 AMD ATOMBIOS (ASCII)
00 B7 28 39 10 00 CRC_Block
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

---

So vbios has two checksums:
CRC-8 for check spi rom for faulty. If this value is Ok - then Gpu core will load firmware
CRC-32 is based in PCI config, and probably checking data and command tables for checksum ok.

Also i compared this value with 4 gpus

Zitat

---

PCI CONFIG CRCs

VAGUHN
02 10 E8 67 00 00 18 00 00 00 00 03 7F 00 32 0F 00 00 00 00 41 4D 44 20 41 54 4F 4D 42 49 4F 53 00
B7 28 39 10
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

PEGATRON
02 10 E8 67 00 00 18 00 00 00 00 03 74 00 32 0F 00 00 00 00 41 4D 44 20 41 54 4F 4D 42 49 4F 53 00
79 01 BC 40
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

DELL 4150
02 10 E8 67 00 00 18 00 00 00 00 03 76 00 32 0F 00 00 00 00 41 4D 44 20 41 54 4F 4D 42 49 4F 53 00
9C 45 55 0D
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00


DELL 4130
02 10 E8 67 00 00 18 00 00 00 00 03 73 00 32 0F 00 00 00 00 41 4D 44 20 41 54 4F 4D 42 49 4F 53 00
EB FE 56 86
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00


Apple Falcon 560X
02 10 EF 67 00 00 18 00 00 00 80 03 74 00 32 0F 00 80 00 00 41 4D 44 20 41 54 4F 4D 42 49 4F 53 00
BF 57 F4 2D
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

---

If you know the location of signature in vbios or driver, or how to bypass this on linux-windows, it could be very useful, thanks!

@bigguygeo as far as I am aware Polaris vBIOSes do not have signatures but the driver does do some kind of check for all Polaris cards (except RX580) where if it finds a modified vBIOS the driver will install but not load. Cleanest solution to that problem is to just make the driver think you have an RX580, the other method is to use a software hack that'll allow the driver to load regardless bypassing the check. I have no idea why AMD put the check into the driver but left out the RX580 in that check, considering Polaris is only any good once you've optimised the GPU frequency to voltage curve and fixed the god awful memory timings it was obviously in AMDs best interest to not add a driver level check at all.

I read some forums research, and before pixelclock patch released (possibly Windows registers patch for bios signature verification), users modded driver by reverse engineering to patch return success from verification stack. So I think amd using crc repository for vbios verification. My suggestion is that: scenario is simply compare. Driver read vbios, generating custom checksum to vbios, reading crc repository and then compare generated crc register with checksum repository and return success, jump not equal to other stack.

So if somebody could reverse checksum generation method, than we can fix by patching one reserved for future data or command module and other info type part of vbios, so it will return success to driver and that’s all.

I wouldn't know on that front @biguygeo you'd need to ask someone who has messed around with the AMD Windows driver but finding someone that has will be difficult nobody wants to touch AMDs Windows driver because from all reports I've seen that driver is a total clusterfuck.