*Official* NBR Desktop Overclocker's Lounge [laptop owners welcome, too]

Nah, don't worry about it. Power limit won't hurt it. I am pulling more than that with my XC2 with the shunt mod and it's totally fine. I am going to try this though, because 1106mV is not possible with any other vBIOS I have tested. That might get me past my brick wall on core clock with the mickey mouse 1.093V cap. Maybe that will finally reverse my GPU voltage vasectomy.

 

I had to use Precision XC1 to see 1106mV but it's definitely there.

 

@Talon - what page on XDEVS is that vBIOS linked to? I am trying to find it. What card is it for?

 

https://xdevs.com/guide/2080ti_kpe/

search XOC vBIOS -- It's a KPE vBIOS for team XOC.

It pulls higher than 1106mV. I saw up to 1125mv with a 100Mhz offset to core. I just started experimenting with it about an hour ago. This might be the vBIOS you need @Mr. Fox

 

Let's hope so. Gonna flash it now and see if it is any better. Thanks, bro.

 

excited! feeling for ya fox! lets see that 1.094+ core voltage!

Sent from my Xiaomi Mi Max 2 (Oxygen) using Tapatalk

 

O it's pulling a lot more than that. I just ran Firestrike and GPU-Z recorded 1.118v at just 30% power limit and 100mhz offset on my ****ty XC card. This could be a long night for me after I swap back to my FTW3 card and flash this vbios to it.

 

Sweet. I did not even apply any overclock or voltage offset and look! 1125mV default. Hallelujah! Now, let me see if I can tune it for a higher overclock on chilled water.

 

@Mr. Fox MayDay >>> best day ever!

 

Good stuff.. this vbios doesn't have the warning sensor readings yea?

Edit:
Ah ok I see it...

XOC BIOS have next limitations to restrict it’s use only for benchmarking purposes:

• All fans are 100% speed (“jet engine mode”). Fan speed curves or adjustment in PX1 is disabled.
• Temp target is removed.
• Power limit is not there anymore.

So the only bios that has the temp sensors reading/protection removed is the LN2 bios eh...

Very nice! Hoorahhh... looking forward to seeing the graphics scores...

I'll be here with ya mate as I'll be benching all night, although I really want to get the desktop together... so might do that inbetween benches.

Also, I haven't even cracked into these yet. I'd like to test the FTW3 w/ the Sammy chips first then swap in the KP card to compare... might not be able to do this all tonight, but I'll get started.



@Rage Set come over for the bbq and we can 4x SLI these lol...

 

Let's see if it actually works right. It seems like it may hold some promise, but I don't want to get too excited yet. I've tested every vBIOS available and pretty much all of them suck (some way more than others).

 

All good. Only one way to find out. Thanks for testing...

 

This is going to take a little trial and error to find out if it is better or not in terms of benching. No access to voltage curve tool, so I cannot lock the voltage where I want it. That sucks. It's going to possibly depend on what the GPU does on its own if I cannot lock the voltage. But, I am getting higher voltage and pulling about 40W more than before. If I do not max out the voltage slider, I cannot achieve my previous max overclock because the voltage doesn't stay locked and the driver TDRs.

I might have to spend more time tinkering with it tomorrow. But, I do like the fact that it is pulling more watts and the voltage is going higher than ever before. If the benchmark scores follow suit it will be good. I have a feeling that it is going to work better stock, but maybe not with an overclock since there is no way to lock the voltage at a fixed value. Dynamic behavior is always bad for overclocking.

 

Sounds very promising. The fact that the voltage slider is even cooperating on max is a good sign (referring to stock vbios w/o shunt mod). I guess you may have to fly manually with maybe a kill-a-watt meter next to you to gauge it the best you can eh?

This is great... so we'll have data from your end with the XOC vbios on the FTW3 along with all of our data with the FTW3 card with both variants of the memory chips... So far the FTW3 cards are preferming very well with 40600+ Graphics in respect to the cooling method. I'll be running on water as well and I think @Talon and @Rage Set set are too ... Also with @Johnksss data... we should be pretty golden with all the data that we need....

 

Yup, I think the inability to lock the voltage is going to be a problem. It runs great until the GPU decides it needs less voltage, allows it to go down just a little bit and then *boom* it's over. Will test more tomorrow.

 

Well, the voltage slider worked for me before with the shunt mod, so that part hasn't changed. It would not stay pegged at 1.093V before except with the voltage slider maxed out. The only problem was the 1.093V cap. I didn't need any higher power limit or anything except more voltage. The higher wattage now is only because of the increased voltage.

 

Ah yes, I should have specified in referring to the cards without the shunt mod and on stock vbios.

I'm really curious now on the graphics score once you get it humming.

 

If I do nothing with overclocking, it stays over 1.100V running with default clocks. So, stock performance will likely be better. Just got to try and figure out if there is some kind of trick to get it to hold the same or higher voltage with an overclock. It really sucks that the MSI Afterburner voltage curve tool doesn't function with this vBIOS. Had the same problem with the Strix XOC 2080 Ti vBIOS. It was totally worthless with that lack of functionality.

Edit: default/stock is 2085 @ 1.125V and never goes below 1.113V while running 3DMark 11. That is with no changes whatsoever using Precision X1 or MSI Afterburner.

 

This is "stock" running on chilled water (radiators not bypassed) with a max temp of 24°C. Minimum voltage was 1.113V under load.

 

This vBIOS put me in the 10K club on Port Royale.

https://www.3dmark.com/3dm/35863288

 

Nice! Congrats, bro.

Yes, the voltage slider is working. But, it seems very sensitive. If you go up too much it actually lowers the voltage and causes the GPU to crash. Maxing out the voltage slider causes instability, so it is definitely working.

It is going to take more trial and error to match or beat my 2220-2235 max core overclock. Now I am starting to think that maybe voltage control may have actually been working before and the readings were never showing over 1.093V with the other firmware I tested. According to XDEVS, this vBIOS fixes the readings with software tools, and that appears to be the case even though I am not using a K|INGP|N card.

The verdict is still out on whether I will be able to overclock any higher, but it is looking promising. I am reserving excitement until that is established, but it's nice to see more than 1.093V whether I can go higher or not. It has been really discouraging to be voltage capped and it makes it easy to hate NVIDIA because of it.

All of these runs are 25 to 50W higher than I was seeing before due to the increased voltage. So far I am seeing a max of 1.125V under load, but it adjusts dynamically and scales with the core clock. (I do not like that part at all. The lag between 2D and boosting to 3D clocks and voltage sometimes causes a crash, whereas it never crashes with the core clock and voltage values fixed/static using the Aorus vBIOS.)

2175 core (+100) and no voltage offset:

2195 core (+115) and no voltage offset:

2200 core (+122) and +25mV offset (TDR crash occurs every time without +25mV offset):

@Johnksss should try this vBIOS if he hasn't already.

Time for bed. We will see what tomorrow brings with the radiators bypassed and the GPU closer to 0°C.

 

Super! Good job gents... awesome...

 

WOOT! Look at this power limit using NVIDIA SMI. This command lowered it from 2000W to 1000W, so K|INGP|N definitely removed the power limit. LOL!!!

@Prema @Talon @Rage Set @Johnksss @Papusan

 

WIDE OPEN THROTTLE.... HERE WE GOOOOOOO....

 

Yup. 2000W power limit per NVIDIA SMI. Man, I am really bummed out that I cannot buy one of these GPUs now.

 

Now time to dunk that in mineral water into the arctic.

In reply to your comment in rep+ - "All good brother... I'm really curious what our FTW3's can pull, because I really think that it can give the KP a run
for its money in the non XOC class... I'm actually more excited to test this FTW3 w/ the Sammy chip over the KP. I'm pretty optimistic that the FTW3 cards can hang....
"

 

Well fellas for the time being the A51M gets to stand in the middle of the podium 1st place as the first fastest recorded in other tests... for what it's worth lol ...

Fire Strike Extreme: 14135

https://www.3dmark.com/fs/19229338

Fire Strike Ultra: 7179

https://www.3dmark.com/fs/19229333

Night Raid: 67195

https://www.3dmark.com/nr/110761

Ok enough with the A51M... now time to go put the desktop back together... I really want to bench this FTW3 w/ Sammy chips...

 

Don't let Umar see that...#TwitterHappy

 

Haha yea you're the first to see it and I'm just leaving it here... I'm much more excited to be posting my FTW3 Sammy chipped runs along with the KP when I get to it.

 

I'll load it up and give it a go. Thanks.

Yes, I had decided to give it a few runs since each time it kept going up. LOL

10/4 negative ghost rider. I seen them scores when I was testing out hybrid sli. And was like, look at @iunlock getting it in. Nice!

 

Oh my it just dawned on me that you're running SLI! Monsterous ... Give those desktop 2080's a run for the money!

 

Sorry, mistake

 

Maybe. They have true Nvlink-SLI, while we have...errr....ahhh....Can I be SLI.

@Talon @Mr. Fox
Max watts i'm seeing is 320W from the card, not the wall. And a max voltage of 1.125V with worse scores....

 

Time Spy: 11664
(The GPU is pulling strong, but the CPU is wonky as heck, yet it pulled a very decent overall score carried by the strong GPU. Work in progress...)

https://www.3dmark.com/spy/7090071

Time Spy Extreme: 5328
(Pretty impressed with the overall here on the Area51M... not too shabby...)

https://www.3dmark.com/spy/7090064

 

Hey Bro, I'm really looking forward to seeing those 2080 Ti KP in action. Congrats to you and Brother@ Johnksss

 

Thank you @JoeT44. I'm really looking forward to it as well. Benching while trying to work on stuff turned out to be pretty inefficient haha so I'll need to just focus on the desktop today. Glad that the weekend is here... First up is the 2080Ti FTW3 Samsung card then onto the KP... hoping to fit all of this in this weekend.

Nice scores to you as well!

 

@iunlock
Your next hurdle, young jedi master.
https://www.3dmark.com/3dm/35873520


Also waiting on the FTW results as well.

 

All green Nice score.

 

This is exactly why i say these cards are pretty much all the same. When one does not look at the score, but only the fps...One can see the real score.

Edit:
Until the KING OF MODS ( @Prema) gets out of his lazy boy 3000 chair.

 

Yeah, same here. I cannot overclock as high (so far) as I could with the Aorus vBIOS. It's equal or better up to 2200MHz on core for me, but going the next bin up on the core overclock never completes. The GPU resets and it crashes in 3DMark 11 Test #2. I may try playing with the power limit (reducing it) and see what happens. And, I don't think it is temperature related. I am getting things colder than I needed to before for 2220-2235 on core.



But, I am seeing what looks like minor performance improvement at 2200 on core with more voltage. Just haven't figured out how to match/beat my previous max overclock yet. Seems like it might be more effort than it's worth.

 

Super! I'll have to make some adjustments to the sail now for TS. As you can see super wonky cpu score on that run...

As for the 50x run with the FTW, coming up soon. I'll be working on the desktop later today.

Broken cpu score. Did you not see? I'm surprised it even pulled the overall score that it did.

Very true... From a fps standpoint almost identical. This 2100 wall needs to be breached lol...

Sounds good and looking forward to it.

 

And I heard people here are having problems with spelling "12K":

https://www.3dmark.com/spy/7090986

https://www.3dmark.com/compare/spy/7090986/spy/7090452/spy/7090071

Now let me get back to my weekend-couch...

 

I commented the scores. All green. Equal the trees here at home that just got green leaves We already know the CPU is wonky in the Area-51m.

 

Hold'em up there speed racer.
https://www.3dmark.com/compare/spy/7091158/spy/7090986/spy/7090071

Why you always raining on the other guys parade.

 

Ha ha ha, I just like rounding up numbers.
There is little in me that even wants to bench something that had already gone over its platou just because others have a higher score. But if I see us jumping around 9.xx while 10 is well within reach, then I really do enjoy giving it a little extra push.

Doesn't really have much to do with other people. Like this one I kept pushing myself against no-one to break 60K and didn't even post it until much later:
https://www.3dmark.com/nr/59081

Guess I just like clean numbers...

Speaking of which, eagerly waiting to see your 35K FS & 20K TS 'SLI'.

 

Interesting. Seems like the power limit slider is working, too. And, pulling that back is bringing a slight performance bump.


@Talon - I don't think I will be able to join the Port Royal 10K club with Micron memory because it won't overclock high enough to get there. But, I think this might be a personal best. I'll need to check to confirm that.

 

Getting the GPU colder doesn't seem to be helping much. Voltage and power limit sliders both work and one affects the other. Looks like this vBIOS works much better up to 2200 on core and then quickly falls on its face above 2200 on core. Maybe it is exceeding the capacity of the XC2 power delivery system, or my shunt mod not playing nice with this vBIOS, or something along those lines. It looks like I am not going to match or beat most of my benchmarks with Aorus Extreme vBIOS. +25mV voltage offset and +25 power limit seem to produce the best results up to 2200 on core. Maybe it's time to find another thing to spend my time and money on. This feels like it is becoming way too complicated and expensive to be a rewarding experience any more. I wish the NVIDIOTS would just stop screwing us over and let their GPUs overclock using the same principles as CPU overclocking like they used to. I am just so fed up with the Green Goblin's idiotic BS where it concerns overclocking. They're too busy playing shill-master and making worthless gimmicks to sell to the stupid little gamer punks hanging out on Twitch and FacePoot.

 

Let´s add this "utilization-block" to "clock-block v3" (2100Mhz) and the list of "strange things that only happen with RTX":