kingx11

bcavnaugh

kingx11

tresnugget
Anyone tried the new bios yet? 

i'm using it now , i think the older Bios was better for OC , i'll keep trying though maybe i got it wrong but so far getting to 2190 is kinda hard

Are you also using the Classified Tool?

yes
1.2v on the core , 1.6v on the memory

Are you going to Start to Post your Scores from 3DMarks?
It would be great to be able to compare them.

i did post FSU in the last page


Thanks I Missed it https://www.3dmark.com/compare/fs/19157501/fs/19051091
You may want to use the Standalone Version over the Steam 2.8.6546 s64 Version to remove Steams Overhead.
Alternatively, you can download the complete installer here.
Remember to write all your Keys down before you change over to the Standalone Version, they are the same.
You can find them under Options in the 3DMark Application.
 

I agree. Could be seeing negative scaling. Also 1.6v mem is pretty high for ambient OC'ing.

When I pushed up to 1.156v Core, I had negative scaling if using LLC 0-3. Using 5 drops temps a bit to stabilize my runs. And for memory, going well over 1.5v just added heat and scaled worse.

EDIT: meant 1.156v NVVDD not 1.56.

I see a lot of people using llc of 3 when the default is 10. Can someone explain why you'd wanna go down. I haven't been experimenting with llc yet but I'm just curious. Even at 10 I'm getting vdroop when pushing higher voltages. Is it that the lower llc, the less vdroop with GPUs vs Mobo/CPU? I'm pretty new at this level of gpu overclocking (as if it wasn't apparent)

kingx11
ok
 
FSU , 9666
 
https://www.3dmark.com/3dm/35655835?
 
how good is that ?!

Still Seems Low 9887
https://www.3dmark.com/compare/fs/19157501/fs/19051706/fs/19051091
Run the Plain Fire Strike and Time Spy and post.
If you have 3DMark 11 run the Performance Benchmark and also post back. 

tresnuggetI see a lot of people using llc of 3 when the default is 10. Can someone explain why you'd wanna go down. I haven't been experimenting with llc yet but I'm just curious. Even at 10 I'm getting vdroop when pushing higher voltages. Is it that the lower llc, the less vdroop with GPUs vs Mobo/CPU? I'm pretty new at this level of gpu overclocking (as if it wasn't apparent)

 
The lower the number, the less VDroop or LLC you have. Basically inverse Asus LLC. 
 
With around Level 0-2 there is much less Droop compared to Level 10. BUT there is more heat and more Power Target used. There is also typically more Ringing in the Voltage line that can be overshoots and transient instability. Basically you'd need an O-Scope with a fast triggering speed to capture the resulting fluctuations. Most DMM (Digital MultiMeter) will not capture the fast transients. 
 
If setting say 1.150v NVVDD: you'd idle roughly the same. But under load, you would maybe catch a spike up to 1.18v-1.20v before it settles to say 1.085-1.105v under load. Once the load completes, the voltage line will rebound. The excess current needs to go to ground. You could see a rebound spike on the Voltage line. 
 
So it's not best practice to just max out LLC and hope it works well. Better practice is to go in steps with NVVDD locked in place and experiment with the LLC levels to see how it effects your OC.
 
Even with a VRM this well made: there's still the laws of physics to contend with.

richiec77

tresnuggetI see a lot of people using llc of 3 when the default is 10. Can someone explain why you'd wanna go down. I haven't been experimenting with llc yet but I'm just curious. Even at 10 I'm getting vdroop when pushing higher voltages. Is it that the lower llc, the less vdroop with GPUs vs Mobo/CPU? I'm pretty new at this level of gpu overclocking (as if it wasn't apparent)

 
The lower the number, the less VDroop or LLC you have. Basically inverse Asus LLC. 
 
With around Level 0-2 there is much less Droop compared to Level 10. BUT there is more heat and more Power Target used. There is also typically more Ringing in the Voltage line that can be overshoots and transient instability. Basically you'd need an O-Scope with a fast triggering speed to capture the resulting fluctuations. Most DMM (Digital MultiMeter) will not capture the fast transients. 
 
If setting say 1.150v NVVDD: you'd idle roughly the same. But under load, you would maybe catch a spike up to 1.18v-1.20v before it settles to say 1.085-1.105v under load. Once the load completes, the voltage line will rebound. The excess current needs to go to ground. You could see a rebound spike on the Voltage line. 
 
So it's not best practice to just max out LLC and hope it works well. Better practice is to go in steps with NVVDD locked in place and experiment with the LLC levels to see how it effects your OC.
 
Even with a VRM this well made: there's still the laws of physics to contend with.

I also really appreciate the instructions as I'd never had tried that. I would've gone the longer way around and started with llc then jumped around on voltage. Your way makes a lot more sense, though. 

TriOx

Spoiler

bcavnaugh

kingx11

bcavnaugh

kingx11

bcavnaugh

kingx11

tresnugget
Anyone tried the new bios yet? 

i'm using it now , i think the older Bios was better for OC , i'll keep trying though maybe i got it wrong but so far getting to 2190 is kinda hard

Are you also using the Classified Tool?

yes
1.2v on the core , 1.6v on the memory

Are you going to Start to Post your Scores from 3DMarks?
It would be great to be able to compare them.

i did post FSU in the last page

Thanks I Missed it https://www.3dmark.com/compare/fs/19157501/fs/19051091
You may want to use the Standalone Version over the Steam 2.8.6546 s64 Version to remove Steams Overhead.
Alternatively, you can download the complete installer here.
Remember to write all your Keys down before you change over to the Standalone Version, they are the same.
You can find them under Options in the 3DMark Application.
 

close

Hey bcavnaugh, looks like you got me off the line with the extra cores but I almost caught up on the top end. 
https://www.3dmark.com/compare/fs/19051091/fs/19198368

Nice, I will be running more latter today. It seems I am stuck at 5000MHz on my CPU.
Looks like this is my max for 3DMarks 11 https://www.3dmark.com/3dm11/13284536 HOF#40
A hard Benchmark to get the right settings.

TriOx

Nice, I will be running more latter today. It seems I am stuck at 5000MHz on my CPU.
Looks like this is my max for 3DMarks 11 https://www.3dmark.com/3dm11/13284536 HOF#40
A hard Benchmark to get the right settings.

To your point, I lost a day to 3dmark11 tuning haha.  I'll probably try it again tonight but doubt I can get much more.
https://www.3dmark.com/compare/3dm11/13284536/3dm11/13281246

 
It took me the last two days to get P38110.

TriOx

 
It took me the last two days to get P38110.

The 9900K/Z390 is way easier. Actually now that you mention it, I'm two days also.  The first was on my i7-9800X/X299 Dark but I've blocked that experience out haha.  I was comparing my X299 scores to yours, last week, and realized I must have a cpu cooling problem so I switched the TIM out and re-seated the water block and managed to get the i7-9800X solid at 5.1Ghz with HT on.  Even after that I was still in the dust.
 
https://www.3dmark.com/compare/3dm11/13284536/3dm11/13266953

1.0.132.0 s is the Steam Version. You might want to change over to the Standalone version to remove Steams Overhead.
Alternatively, you can download the complete installer here.
Remember to write all your Keys down before you change over to the Standalone Version, they are the same.
You can find them under Options in the 3DMark Application.

Stupid question time..

When its said to restart to enable the changes.. does that mean to restart the pc or the classified program?

Also the Loadline setting.. I dont have the proper hardware to test.. higher number more droop, lower equals less?

AHowes
Stupid question time..

When its said to restart to enable the changes.. does that mean to restart the pc or the classified program?

Also the Loadline setting.. I dont have the proper hardware to test.. higher number more droop, lower equals less?

The Computer.
Using Loadline of 10 See https://xdevs.com/guide/2080ti_kpe Loadline was added mid way down.

Thanks!

One last one for now.. if one is using ab and classified tools and sets over 1.093v like 1.1+ and it still only shows 1.093v on the osd, is it using 1.093v or really 1.1v and it just cant display the correct voltage.

Edit:
Ab only shows what voltage you set in there software. I believe px1 shows the actual voltage used.

Seeing this error on the new Classified Tool.

TiN_EE
Try again? Maybe download failed? :)

I did 8 times now from 2 different computer but will give it another go.
Still error as above under Windows 7 and Windows 10.