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

It looks like NVIDIA finally sold all of their defective Turing chips to early adopters and now they can spin this marketing stunt as if they are doing something special for customers by selling them normal parts for the same price as the defective parts that were sold to suckers for normal prices, LOL.

NVIDIA Rolling Out More Powerful TU106-410 And TU104-410 Dies At The Same Price For 2070 And 2080 Graphics Cards

But, the NVIDIOTS aren't the only ones selling gimmicks. Paying extra to get absolutely nothing in return seems to be a popular trend in PC tech.

11:38 - "Nothing has changed. AMD's processors are at the limit already." "Don't buy this for performance, because that is not what you are getting."

 

I flashed back my stock EVGA XC2 vBIOS for comparison and the K|INGP|N XOC vBIOS pulls right at 100W more power from the wall in 3DMark 11 test #1 with the core clocked at 2200MHz.

 

Well... Actually, this is not accurate. I was mistaken. I am pulling as much or even more watts from the wall with the stock XC2 Ultra vBIOS. The shunt mod clearly worked much better than I thought it did, as the K|INGP|N firmware doesn't increase the power limit at a hardware level... basically the same outcome except the hardware mod defeats the cancer firmware and the K|INGP|N firmware has no cancer that needs to be defeated. And, I am thinking that I am applying voltage I never knew I had because the reading maxes out at 1.093V with the cancer firmware. I am seeing zero benefit to the K|INGP|N XOC firmware and I'd say it is working really well. If it eliminates the need for a mod involving soldering, I'd say that is a better solution even though I am not going to really see any benefit in terms of benchmark scores.

The part that has me stumped is that using the stock XC2 firmware, FTW3 firmware and the Aorus Extreme firmware I have no performance cap reasons, but all of the "enthusiast" firmware has voltage and power performance limit reasons, and it is no exception to that rule with the K|INGP|N firmware.

I'll upload a video to YouTube later tonight or tomorrow to show exactly what I am talking about.

 

I’m finding the vBIOS for me works fairly well in terms of giving me far more performance in benches I was previously power limited. I was hitting power limited results with my stock FTW3 vBIOS. The XOC vBIOS for me is cancer free, just missing a voltage curve/lock. Still I never found that as useful for my own needs and it didn't always help me score higher just by locking in a voltage (probably because I was power limited). I agree with you that it definitely removes the need to do the hardware mod/shunt and I can just use this vBIOS for my benching and testing, then flip the bios switch back to normal vbios for daily use or gaming.

The bios seems to react better when I lower the power slider down and don’t touch the voltage. I haven’t had much time to mess around with it today but will start again soon. I think @Mr. Fox You might be hitting that 2200mhz wall that many KPE cards are also hitting.

Were also seeing the issues the KPE guys are seeing, except they have the XOC classified tools to mess with the voltage more finely, tune LLC and other fun things. Been reading their posts and it seem like a lot of work and trial and error to see what works for your particular card. If we had access to those tools it would make it even better.

 

Yeah, I think this does eliminate the need for the shunt mod.

You game a whole lot more than I do, and none of this effort is needed for gaming. And, yeah... That 2200 core clock limit is ridiculous. I can get it stable at 2220 with +75 offset on core voltage, +25 stable for 2200. For it to happen on a bleeding edge GPU is totally inexcusable. If I paid $2000 for a K|INGP|N card and it performed no better than my XC2 I would be so severely pissed that I'd be beside myself. I'm thinking about selling this GPU and replacing it with a FTW3 if I can be assured it will have Samsung memory, but part of me just wants to be done with the nonstop flow of broken feces that runs like crap out of the box and needs special firmware and/or hardware mods to achieve full performance.

I've learned a lot more about my GPU using this K|INGP|N XOC firmware because I can actually see more clearly what is happening now, and it wasn't as evident with false sensor readings. I think you'll like my video. It's almost done.

 

Indeed!

Funny you should mention that tool. As of right now it seems to not work with the XOC vbios... Works with the 3 factory files that come installed on the card though.

 

Saw that coming (12K) Great score!

Two thumbs up!

I've been noticing the nice rounded numbers haha. As for rounding up yea it's true..when you're so close reaching for going over the next number, it does make you want to get over that hill...

Very interesting and great info...thanks for sharing...

The Great Wall of 2100MHz. RTX (Restricted - Turing - X-Factor)

 

As soon as the video processing is complete and it is viewable in better than 360p resolution, I will click on "publish" and post a link to it here. The upload finished about 30 minutes ago, so it shouldn't be too much longer.

 

Sounds great... I'll play it on the tablet next to me, while I tackle getting this desktop back together to do some FTW3 runs.

 

OK, wow, that took a long time to finish processing on YouTube's end. I included a poll on this one near the end of the video. I did not include any narration, but it should (hopefully) be obvious what I am demonstrating. @Talon - still not enough voltage. But, the voltage slider is working and so is the power slider. Setting the power slider to 25% set 500W and when I used NVIDIA SMI to move it to 2000W the dialogue in the command window said exactly that.


@Johnksss

 

Sweet...watching now... thanks for taking the time to put this together.

Edit: Just watched it...great vid and great job!

 

5Ghz

AW CPU performance getting wrecked. I must get to the bottom of this. Hoping to put in a good hour playing with my Area-51m finally

I've noticed this behavior too a few times.

 

Yea it's frustrating ... I've been benching all night trying to figure out the cpu wall... such valuable time I could be spending on doing other projects lol.

 

Thanks Brother @Mr. Fox for the time and effort you put into making that video.

 

@Johnksss

This has "70K trap" written all over it:

https://www.3dmark.com/nr/111500

Lord give me the strength to resist!

 

Good morning fellas... @Johnksss I've got the desktop all put together last night so I'll be diving into the FTW later today. The A51M soaked up so much of my time the past few days lol I'm leaving it to rest so that I can focus on other things. Great Night Raid score btw... break that 70K!

@Mr. Fox ... we shall see eh ... If the graphics score on the FTW card are within paper football flicking distance to the KP..... :X lol ... (Edit to avoid confusion, meaning on air/water that is...not XOC or anything in that level.)

 

As to the kingpin card not working...I beg to differ... And he is under 2200Mhz...
https://www.3dmark.com/fs/19235863

 

So, Samsung is killing off B-Die production, Nvidia is killing off product of A and non-A variants of their Turing GPU's (at least for 2070/2080). AMD may not be releasing TR 3000 any time soon. Looks like the market is focusing on the high to med mainstream parts. Death of the HEDT?

 

Who said it wasn't working?

Seems like everything awesome ends up getting killed at the hands of its creator. The mediocrity, ignorance and low standards of the mainstream masses ends up screwing us all. The concept of good being the enemy of great has been lost to the popularity of everyone, including losers, getting a trophy for participation and feeling "included" when they haven't earned it.

Until we get back to the basics and start calling the balls and strikes in a militant and antagonistic fashion, not allowing what is wrong to be viewed as "OK" and making inferiority undesirable (like it should be) this is only going to get worse. It is ruining things in other areas of life, too. It is totally screwing up corporate work environments... making it difficult to differentiate high performance employees from half-assed deadbeats based on rating/ranking, performance evaluations and compensation. No matter where it surfaces, this always sucks.

 

Was referring to the FTW3 cards being on par with a working kingpin card.
Also to add no card is going above 2250 on water unless one is really really lucky. These cards run just like the cpus. If you are not in negative temps, it will not gain huge leaps in performance.

 

Ah, OK. I see you mean now. I thought maybe there was a typo or error in something I posted. The firmware alone makes an ordinary card like mine run better because the ordinary firmware is a screwed up mess. There is no question in my mind that the K|INGP|N cards work correctly and represent the best the industry has to offer.

I think what @iunlock was referring to is running an FTW3 on air, water or chilled water might end up producing benchmark scores that are not remarkably different from a K|INGP|N card operating under the same conditions because of how the screwed up and crippled the behavior of NVIDIA products has become. I don't know that he meant the FTW3 cards were as good or have the same potential under extraordinary operating conditions. They clearly do not have the same potential where extreme cooling is concerned because they lack many of the tuning capabilities.

I still think it sucks that things have been compromised to the extent that they have. When a GPU cannot go over a certain clock speed when it is still 50°C below a temperature that qualifies as being "too warm" and begins to throttle only because the firmware is garbage, that really sucks. It also diminishes the already poor value of the best products to have to run them on LN2 to see a substantial improvement in how they perform compared to the ordinary products. If you pay 25% more for a better product, it should provide at least 25% better performance out of the box, and without having to use extreme measures to identify an advantage to buying it. I think there is no excuse for how NVIDIA has butchered things. Even people that don't care about overclocking pay big bucks for products that throttle and malfunction at normal operating temperature. That sucks.

 

This is pretty much true. And i agree on just about all of it except for the LN2 part. No way i'm going for people on water to have the same benefits as me using phase change, dry ice or LN2. I'm never going to agree with that. It's a pay to play game and right now i'm kind of lazy in that i'm not running my SS or using my LN2 pots.

 

Sooo, I installed my new dual DDC pump into the CPU loop and I finally got these KP cards installed. Almost a full EVGA build.

 

I stand corrected.
2280 on core and partial waterblock
https://www.3dmark.com/fs/19051176

 

You're 100% correct. I am not saying anyone should have the same benefits that paid less. I'd be totally pissed off if that were true, but the reality is is sort of seems like it is true in some ways. And, it should not be true under any circumstances.

What I am saying is your K|INGP|N cards should not have to be run on LN2, DICE or phase change cooling in order to have remarkably higher benchmark scores and be more enjoyable to use than a card that costs 25% less. You should see 25% better results under all operating conditions because you paid more and should get more no matter how you are cooling it. The fact that all of the "better" 2080 Ti GPUs perform almost exactly the same (excluding those cursed with Micron memory clock limits) and have more or less the same max core clock running on air, water or chilled water is totally screwed up. I suspect that is the part you agree with.

Looking at it another way, I am really bummed out that I cannot buy a K|INGP|N 2080 Ti without making my wife upset. I have no plans on doing any LN2 cooling. I might someday, but probably not due to the added cost involved. If I can buy a FTW3 for $500 less (which would still piss my wife off) and see more or less the same benchmark scores on chilled water, or only a tiny bit better with the K|INGP|N GPU, then there is something messed up about that scenario... if that turns out to be accurate. Not saying it is accurate, only that I would be unhappy about it if that turned out to be accurate if I paid the extra premium to buy the best money can buy. But, it kind of seems like it might be the case. And, I blame NVIDIA for that, not EVGA or Vince. They're just doing the best they can with what they have to work with.

There also should be very little silicon lottery variance between GPUs that cost more. People that go the extra mile and pay a lot extra for the better product should expect them all to run almost exactly the same and not have to hope to get lucky in the silicon lottery after busting open their wallet for the best money can buy. And, there should be some sort of guarantee that you are going to get a minimum of X% performance increase for the super-expensive premium GPU to not be deemed defective by the manufacturer and eligible for RMA.

 

That's why I post the link earlier and the one right above your post. The gpu scores are higher. Not sure about getting a 25% increase in anything at the moment. That seems to be a lot to ask for given nvidia's current track record.

 

Looking good, Brother @Rage Set

 

I will seriously have to re look at this card.
https://www.3dmark.com/fs/19051394

Could also be because of the high over clock on the 9900K helping I suspect.

 

What EVGA GPU model is he running? I don't see where 3DMark identifies one GPU model from another.

It could be due to the higher CPU clock, but the graphics score is only about 5-6% higher. If he paid a whole lot more money for his GPU and is using extreme cooling to get a 5-6% bump, that's not much of an increase in performance for the amount of money and effort it took to achieve it. If it were 5-6% better performance running under normal conditions (air, ambient water, etc) and scaled higher with the extreme cooling effort and costs, then it would be a nice return for the added cost.

https://www.3dmark.com/compare/fs/19051176/fs/19241172

 

That's the same mathematical problem with any GPUs that are already at 2200Mhz, you'd need 2420Mhz to even get close to 10% more performance, or 2640Mhz to get close to 20% more performance (that's if performance even scaled perfectly linearly with clockspeed, which it doesn't in most cases. What I'm saying is that at these high 'stock' clockspeeds that Pascal & Turing are at, then it takes many hundreds of extra Mhz on top of that to make any significant peformance difference from a percentage point of view - an extra 100Mhz at those clockspeeds is basically nothing (less than 5%). I just think these architectures and silicon has just plain run out of clockspeed, they're already at their limit pretty much - I don't think power & temperatures are the limit here. If you think about the massive clockspeed jump from Kepler/Maxwell to Pascal/Turing, that's quite a feat they have pulled, probably not too suprising that there's not much left in the tank.

 

Yup, that is true. They are running out of headroom and you cannot fabricate something that does not exist. But, the fact that it is true further diminishes the value in purchasing premium products. There has to be a point when winning at benching is no longer a worthy pursuit where the unreasonable increase in the purchase price of the main components is concerned. Winning is everything when it comes to benching, but something that at least partially resembles financial equity would be nice.

And, we know that Pascal and Turing GPUs have untapped headroom that is being manipulated to the Green Goblin through cancer firmware so they can make money re-releasing essentially the same product later on without having to make anything new.

While not on the topic of overclocking, but similar in principle, this video is a great example of what scammers NVIDIA have become. Charging exorbitant premiums for something that is close to worthless and delivers no added value. We can apply the same concept to the clock-blocking bull feces scam. The victims are gamers, not overclockers, but they're still getting screwed over.

 

That is a 2080 ti Kingpin card with a partial water block no water chiller.

He is running a partial water block and no chiller as far as I can tell.
The card is a 2080 ti Kingpin
And if it was extreme cooling it would be in the 2400+ range.

 

With 1000+ to the memory and 100+ to the core, I have already hit within striking distance of my top Time Spy Extreme score produced with my Asus 2080 TI and EVGA 2080 TI FTW3 Ultra Hybrid.

This is going to be fun!!!

 

Links?

 

https://www.3dmark.com/compare/spy/7102664/spy/6315197

My first attempt with the KP cards without the XOC bios versus my best in Time Spy Extreme

EDIT: I should say that my FTW3 Ultra Hybrid wasn't the limiting factor, it was the Asus 2080 TI OC. Even with a custom water block, the Asus couldn't keep up.

 

I have seen others that had the same experience with their ASUS GPUs not keeping up. The Strix XOC 1080 Ti was a real beast, but the RTX 2080 Ti Strix XOC seems to not rise to the same level as its predecessor.

How does one FTW3 Ultra Hybrid (SLI disabled) fare against one KP Hybrid (SLI disabled) running the same core and memory offsets on both GPUs? Are you able to test them both, or the FTW3's are no longer installed in a machine?

 

@Mr. Fox
Mine totally sucked, that's for sure. (Asus 2080 ti cards)

Edit: I also would like to know how the FTW3 card is fairing as well.

 

https://www.3dmark.com/compare/fs/19253295/fs/19253437

There you guys go

EDIT: Both have Samsung memory

 

dunno about u guys but to me that difference looks like regular benchmark variance. so basically equal at equal clocks.

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

 

Awesome, thank you.

Yes, you can have more variance than that running back-to-back passes on the same machine without changing anything.

 

So at a fairly decent overclock they hold the same. Thanks.

Looks to be about right, but it's the higher you go that you start to notice things.

 

What if you set +135 on core and +1300 on memory? Will both run without locking up or artifacts?

 

I know my Hybrid can go as high as +145 on the core and +1260 on the memory with the stock AIO. Anything else without better cooling doesn't stick. (That was when it was paired with the Asus card, so I had to balance the clocks with it, so it may be able to go higher).

 

Well, that is definitely better than most. +122 core and +1125 memory is the most I can get with the Aorus Extreme firmware (+135/+1150 with other firmware, as the Aorus base clocks are higher at their defaults). Of course, that is Micron memory, not Samsung.

 

https://www.3dmark.com/3dm/35908720

My FTW3 Hybrid with +100Mhz Core and +1000Mhz memory (will do +1500Mhz) on the XOC vBIOS.

XOC and it's crazy power draw heat my GPU up about 10C higher. I am looking at grabbing the copper block, a rad and pump. I need something that can fit into my Corsair 450D case. The top is occupied by my AIO for CPU and would rather the closed loop only feeds the GPU to keep temps down and maybe get away with a smaller rad. I could do a 280mm rad in the front. Curious what that would do my overall case and CPU temps once I draw the cool outside air over the rad and bring that into my case.

 

Hummm, i think the highest i got to was +165, but not always stable.

 

I think us KP owners have to watch our backs now....we might get outdone soon.

 

I did read something about the falling demand of B-Die, which is sad, but with lower production that will only increase the price. Fortunately companies like G.Skill are using hand picked B-Die for their upper end line of RAM so that's one sure way to get the cream of the crop.

The only exciting thing, yet unknown, on the horizon is the AMD 3000 line of cpu's... I think we've hit a plateau with gpu stuff as the only way for that to improve is for the memory chip makers to advance. We'll see how it all plays out.

Precisely what @Mr. Fox said. That's what I was referring to, not XOC stuff... "I think what @iunlock was referring to is running an FTW3 on air, water or chilled water might end up producing benchmark scores that are not remarkably different from a K|INGP|N card operating under the same conditions because of how the screwed up and crippled the behavior of NVIDIA products has become. I don't know that he meant the FTW3 cards were as good or have the same potential under extraordinary operating conditions. They clearly do not have the same potential where extreme cooling is concerned because they lack many of the tuning capabilities."

Looks awesome!

That's my buddy who I was referring to with the KP a few pages ago. Correct @Johnksss - It's on a partial water block and that's it... pretty strong pulls... Here's his monstrous TS score too on just water: https://www.3dmark.com/spy/6910334

It'd be nice to have an equal ratio of price to performance, but unfortunately that doesn't exist much at all sadly.

It really is a pay to play game and in short IMO (ideally) for the sake of points it's worth much more to get two FTW cards to rack in double the points in several different categories by a factor of 2x and/or however many cpu's you have to play in all the different categories.

There's a lot of truth to that expanding across many aspects of the field.

Under "normal," conditions the real world difference is non-existent, but on the flip side people who buy (pay to play) the top end KP cards are benchers and not gamers.

It really is a novelty item and yes under "normal," circumstances the difference is miniscual at best and a value of zero to a gamer, because there's no way they can distinguish the difference between a few frames or even less.

KP cards, definitely not for gamers IMO lol ... it's for us benchers.

 

Hope everyone had a great Sunday...

Flushing the loop now. Had to redo some parts to accommodate the new reservoir. All good now... About to fire it up soon. The FTW (Sammy) card is in there now....

I found this in my misc. bin on the shelf while looking for my PSU pump adapter lol...


Seems pretty accurate eh...Ok I guess it can stay...

 

EVGA GeForce RTX 2080 Ti XC2 ULTRA GAMING **WITH HYDRO COPPER BLOCK***

Are both of those items in your photos related (part of the same product)? Got a link to it? Looks nice.