[affxct]

Guys, for some reason my card is running at PCIe X16 1.1? What the hell.

Edit: Nvm it’s just power saving. My friend scared me a bit cause his Carbon was sitting at 4.0 X2 for some reason.

 

[Vld]

Just out of curiosity - what is your opinion on "z690 vs z790" motherboards ? Afaik there is little to none difference between those two chipsets, bit more PCI gen5 lanes and one more USB3.2.

 

[affxct]

Getting 1T stable on the Dark feels impossible. Trying to setup a low voltage 7000-1T and it feels like errors arrive at nonsensical times. At 1T errors also seems to come in large bunches.

 

[Stahlstich]

I currently have no problem with 1T. Only above 7200 nothing works.

 

[affxct]

I currently have no problem with 1T. Only above 7200 nothing works.

View attachment 2580629

Wish this were me. I can get 2T OCs working fine. I’m currently testing 7000C34 with low IMC voltages, but somehow 1T just feels like an impossible task. By this point I’m convinced some kits and boards just aren’t as good at running it.

 

[warbucks]

Wish this were me. I can get 2T OCs working fine. I’m currently testing 7000C34 with low IMC voltages, but somehow 1T just feels like an impossible task. By this point I’m convinced some kits and boards just aren’t as good at running it.

Honestly, the difference between 1T and 2T is so tiny. I would focus on 2T and higher frequency.

 

[affxct]

Honestly, the difference between 1T and 2T is so tiny. I would focus on 2T and higher frequency.

I had 7200C32 after a few hours with RPL, but that’s about where M-die stops scaling. I’m not sure if one can call 1.605/1.515 scaling though. I’m currently switching between 7000C34 1.455/1.365 and 7000C32 1.5/1.41 because of the far safer DRAM voltages and 1.25 TX/1.0 SA. I want these sticks to survive as long as possible. 1.45 is probably the play tbh. It seems like vendors are reluctant to bin higher than 1.4V. It’s starting to become mega sus.

 

[z390e]

Anyone have a good thermal pad recommendation for the m.2 slot on Dark?

 

[Avacado]

Anyone have a good thermal pad recommendation for the m.2 slot on Dark?

Amazon.com: Thermalright Thermal Pad 0.5mm 12.8 W/mK, 85x45x0.5mm, Non Conductive Heat Resistance, Odyssey Silicone Thermal Pads for PC PS5 PS4 Laptop Heatsink GPU/CPU/SSD/VRAM/LED Cooler : Office Products

Amazon.com: Thermalright Thermal Pad 0.5mm 12.8 W/mK, 85x45x0.5mm, Non Conductive Heat Resistance, Odyssey Silicone Thermal Pads for PC PS5 PS4 Laptop Heatsink GPU/CPU/SSD/VRAM/LED Cooler : Office Products

www.amazon.com

Amazon.com: Fujipoly/mod/smart Ultra Extreme XR-m Thermal Pad - 60 x 50 x 0.5 - Thermal Conductivity 17.0 W/mK : Electronics

Buy Fujipoly/mod/smart Ultra Extreme XR-m Thermal Pad - 60 x 50 x 0.5 - Thermal Conductivity 17.0 W/mK: Electronics - Amazon.com ✓ FREE DELIVERY possible on eligible purchases

www.amazon.com

TG-PP10-50 t-Global Technology | Fans, Thermal Management | DigiKey

Order today, ships today. TG-PP10-50 – Thermal Silicone Putty 50 gram Container from t-Global Technology. Pricing and Availability on millions of electronic components from Digi-Key Electronics.

www.digikey.com

 

[Agent-A01]

Amazon.com: Thermalright Thermal Pad 0.5mm 12.8 W/mK, 85x45x0.5mm, Non Conductive Heat Resistance, Odyssey Silicone Thermal Pads for PC PS5 PS4 Laptop Heatsink GPU/CPU/SSD/VRAM/LED Cooler : Office Products

Amazon.com: Thermalright Thermal Pad 0.5mm 12.8 W/mK, 85x45x0.5mm, Non Conductive Heat Resistance, Odyssey Silicone Thermal Pads for PC PS5 PS4 Laptop Heatsink GPU/CPU/SSD/VRAM/LED Cooler : Office Products

www.amazon.com

Amazon.com: Fujipoly/mod/smart Ultra Extreme XR-m Thermal Pad - 60 x 50 x 0.5 - Thermal Conductivity 17.0 W/mK : Electronics

Buy Fujipoly/mod/smart Ultra Extreme XR-m Thermal Pad - 60 x 50 x 0.5 - Thermal Conductivity 17.0 W/mK: Electronics - Amazon.com ✓ FREE DELIVERY possible on eligible purchases

www.amazon.com

TG-PP10-50 t-Global Technology | Fans, Thermal Management | DigiKey

Order today, ships today. TG-PP10-50 – Thermal Silicone Putty 50 gram Container from t-Global Technology. Pricing and Availability on millions of electronic components from Digi-Key Electronics.

www.digikey.com

Don't waste your money on those.

Arctic TP-3(new product) is better and way cheaper.

Amazon.com: ARCTIC TP-3: Premium Performance Thermal Pad, 100 x 100 x 1.5 mm (1 Piece) - High Performance, Particularly Soft, Ideal Gap Filler, Bridging Gaps, Safe handling : Everything Else

Amazon.com: ARCTIC TP-3: Premium Performance Thermal Pad, 100 x 100 x 1.5 mm (1 Piece) - High Performance, Particularly Soft, Ideal Gap Filler, Bridging Gaps, Safe handling : Everything Else

www.amazon.com

Plus you don't need anything with a high thermal conductivity for m.2 drives.

 

[Vld]

Got weird issue on Z690 DARK with 12900KS. I did set up OC to all core 54 with avx offset 3. Non avx workloads run great as soon as i run avx/avx2 stress test all P and E cores downclocks to 2.2 Ghz, ring 2200 mhz. If i manually set ring to any number other then 22 - cores jump to correct speed. Bios 2.03. Am i missing something ?

 

[affxct]

The return of reboot instability. I’ve managed to replicate something that I thought I had experienced last week.

I’ve come to find that if one changes settings that significantly impact memory training, and then revert to a profile that was previously stable, training will not be proper and you’ll end up error’ing almost as though you had not reverted your changes. I’m not sure if this is related to me being at the limit of my board/IMC or if TX VDDQ being 1.35 for 7400 is too low (could be this), but clearing CMOS for 30s while the board still has a charge in it seems to work.

Once the board auto restarts and you head back into BIOS to load your profile, training for that profile should be proper and you should be stable once again. The board has to have a charge and it has to auto boot once you leave the CMOS clear button. I’m not sure how long it has to be cleared for, but usually 30-90s works well.

 

[FreeSpeechIsKnowledge]

The return of reboot instability. I’ve managed to replicate something that I thought I had experienced last week.

I’ve come to find that if one changes settings that significantly impact memory training, and then revert to a profile that was previously stable, training will not be proper and you’ll end up error’ing almost as though you had not reverted your changes. I’m not sure if this is related to me being at the limit of my board/IMC or if TX VDDQ being 1.35 for 7400 is too low (could be this), but clearing CMOS for 30s while the board still has a charge in it seems to work.

Once the board auto restarts and you head back into BIOS to load your profile, training for that profile should be proper and you should be stable once again. The board has to have a charge and it has to auto boot once you leave the CMOS clear button. I’m not sure how long it has to be cleared for, but usually 30-90s works well.

That's correct. I can't get past 7400 either. I think it's a bios limit at this point or the boards limit but, I somehow doubt it's a hardware limit.

 

[affxct]

That's correct. I can't get past 7400 either. I think it's a bios limit at this point or the boards limit but, I somehow doubt it's a hardware limit.

I’ve heard and seen people doing more than 7400, so it can’t be a board limit, but this re-training instability worries me.

 

[FreeSpeechIsKnowledge]

I’ve heard and seen people doing more than 7400, so it can’t be a board limit, but this re-training instability worries me.

I just figured out I can indeed run 7600MHz memory. I had to bump up the IMC PLL voltage this whole time... I sat here this morning going through all the voltages going up and down one by one finding the median they liked, and it was failing consistently no matter what. I did balance my other voltages out good though and figured I'd try to set the IMC PLL setting at the bottom of the OC BIOS tab to 1.4 and it's stable now and running tm5 extreme.

Just figured I'd post this info here for people having issues getting over 7400 like I was with their 7600 kit.

Never mind it went unstable again after a reboot.

 

[Muut]

Did you input those voltages manually ? Other than VDD, I have always left everything on Auto with the Dark (can take up to 15min to train but in the end, it always managed to train and boot as long as timings were no incoherent).

Here are some profiles I posted in the DDR5 Daily channel. Again, voltages are set on Auto but I'm reporting what the motherboard has chosen for the given frequency.

It's all m-die so it probably needs more voltages overall than you, but still I find your VDD2 to be a bit low despite being a-die. I have a-die coming on monday, I'll give it a try.

Stable :

7200 c32-42-42-28 (m-die)
SA: 1.1v
VDD : 1.54v
VDDQ : 1.45v
VDD2 (IMC) : 1.54v
VDDQ TX : 1.4v

Benchable profile :


7600 c32-42-43-28 (m-die)
SA: 1.2v
VDD : 1.66v
VDDQ : 1.63v
VDD2 (IMC) : 1.62v
VDDQ TX : 1.4v

Almost same profile but for Y-cruncher.
Had to loosen a bit :

7400 c32-44-44-28 (m-die)
SA: 1.2v
VDD : 1.66v
VDDQ : 1.63v
VDD2 (IMC) : 1.62v
VDDQ TX : 1.4v

8000 C32-44-44-28 (m-die)
SA: 1.35v
VDD : 1.70v
VDDQ : 1.66v
VDD2 (IMC) : 1.64v
VDDQ TX : 1.4v

I just figured out I can indeed run 7600MHz memory. I had to bump up the IMC PLL voltage this whole time... I sat here this morning going through all the voltages going up and down one by one finding the median they liked, and it was failing consistently no matter what. I did balance my other voltages out good though and figured I'd try to set the IMC PLL setting at the bottom of the OC BIOS tab to 1.4 and it's stable now and running tm5 extreme.

Just figured I'd post this info here for people having issues getting over 7400 like I was with their 7600 kit.

View attachment 2585189

Never mind it went unstable again after a reboot.

 

[_AntLionBR_]

.

 

[FreeSpeechIsKnowledge]

Did you input those voltages manually ? Other than VDD, I have always left everything on Auto with the Dark (can take up to 15min to train but in the end, it always managed to train and boot as long as timings were no incoherent).

Here are some profiles I posted in the DDR5 Daily channel. Again, voltages are set on Auto but I'm reporting what the motherboard has chosen for the given frequency.

It's all m-die so it probably needs more voltages overall than you, but still I find your VDD2 to be a bit low despite being a-die. I have a-die coming on monday, I'll give it a try.

Stable :

7200 c32-42-42-28 (m-die)
SA: 1.1v
VDD : 1.54v
VDDQ : 1.45v
VDD2 (IMC) : 1.54v
VDDQ TX : 1.4v


View attachment 2585203



Benchable profile :


7600 c32-42-43-28 (m-die)
SA: 1.2v
VDD : 1.66v
VDDQ : 1.63v
VDD2 (IMC) : 1.62v
VDDQ TX : 1.4v

View attachment 2585204

Almost same profile but for Y-cruncher.
Had to loosen a bit :

7400 c32-44-44-28 (m-die)
SA: 1.2v
VDD : 1.66v
VDDQ : 1.63v
VDD2 (IMC) : 1.62v
VDDQ TX : 1.4v

View attachment 2585205

8000 C32-44-44-28 (m-die)
SA: 1.35v
VDD : 1.70v
VDDQ : 1.66v
VDD2 (IMC) : 1.64v
VDDQ TX : 1.4v

View attachment 2585206

That is super helpful thank you. I'm updating the other two Bios chips to 2.03 right now as well.

I'll report back where I get. Thank you again.

 

[FreeSpeechIsKnowledge]

Can we please get a preview release of BIOS 2.04? I'd appreciate it as I'm stuck at 7400 right now and it's stable as a rock at that and I'd really like to push my setup at this point.

 

[affxct]

Can we please get a preview release of BIOS 2.04? I'd appreciate it as I'm stuck at 7400 right now and it's stable as a rock at that and I'd really like to push my setup at this point.

Microsoft OneDrive - Access files anywhere. Create docs with free Office Online.

Store photos and docs online. Access them from any PC, Mac or phone. Create and work together on Word, Excel or PowerPoint documents.

1drv.ms

It saved to my Google Drive this morning when I clicked on the link on my iPhone XD. They couldn't stop us. I ported the file into the flash tool and renamed it appropriately. Please bare in mind it's a non-EVGA-official file so like, it's totally at one's own risk. I don't quite have time to test right now but let me know how it goes if you do.