Mr. Fox

Just to add one more to the endless list of reasons people should be using NVCleanStall to rip out all of the worthless garbage the Green Goblin bundles in its payload of filth. (A payload of filth is becoming common in all drivers, not only GeFarts.) Why, pray tell, would ANY driver need to include telemetry malware? Short answer: no such need exists. We all get to partake of the stinking garbage that the world's lowest common denominators are willing to put up with. This applies to everything in life. We have many millions of brain-dead losers that submit to the will of wicked decision-makers to thank for everything that sucks in the world around us.

That sucks. I have had that happen to me before, even with all avenues for updates disabled and thought to be broken and non-functional by design. Makes good sense to me. I do think the issue is something not right with the AIO. Whether it is fit, pump operation, debris in the AIO loop, diminished coolant volume, or something else is anyone's guess at this point. If changing the AIO fixes the problem, all of the other enhancements (delid, liquid metal, IHS lapping, etc.) will just be icing on the cake. If it doesn't... well... cross that bridge next.

My impression is that something is not right with the AIO. You will certainly see improved results with a delid and liquid metal if the fit is good, so that is something worth considering regardless of the root cause. I would do that now and see where it leads and it will be out of the way and ultimately provide a benefit even if it doesn't solve the problem you are having. If fit is poor and doesn't make good enough contact for liquid metal, you should see improved temps after delid using thermal paste between the IHS and cold place. I would start with the idea that the AIO is the problem if I were in your shoes, but still do the delid first.

Yes, either term can be used. They are almost the same thing. A scam is premeditated and organized fraud. Fraud can happen when someone sees an opportunity to take advantage of, and does so, even if it is not the result of a thoughtful plan. Both are deliberate. Fraud does not happen by accident.

Did you mod the voltage and power limits using RBE? Or, cross flashing from a different brand? @Rage Setgave me links to read about the tweaking tools and where to download them. I think it will be interesting to experiment with. I wish NVIDIA firmware could be messed with in the same way. Cross flashing them is easy, but they have all kinds of Nazi blocks on modded firmware. I don't know how delicate the Radeon cards are though, or if they are easy to kill/brick, or if they are resilient and can take a lot of abuse. This is uncharted waters for me, so I will have to investigate further before diving in. Yeah, $699 for a 6950XT is reasonable. You can't even find an obsolete 3090 for that price as far as I have seen.

Right. But, being charged for it as if that were true is totally dishonest. Even more dishonest when it performs worse than something that cost $300-$500 less. I agree with that. While the volume of water getting pumped through the block matters, fit is more important. If it cannot transfer the heat off of the die or IHS due to poor contact then running more water through the block will accomplish nothing. The fact that the phase change pad worked better than thermal paste for a period of time, but ultimately failed, suggests there is a problem with fit and/or contact. If the fit is poor enough, liquid metal might not provide any help.

Nope, I did not. Brother @Rage Setis loaning me a 6900 XT and waterblock to play with and if I like it I will buy it instead of a 4090 for now. I can always get a 4090 or 4090 Ti later on, or maybe not ever. I'm not liking how firmware crippled they are in terms of overclocking the core and the lack of voltage control sucks. That does not surprise me, but for the price they are charging you should get more than a gamer-wuss cookie cutter GPU. And, paying more should get you more. In the case of the Strix paying more meant getting less. Well, that's odd. Your link gives me this:

I also texted you my mailing address to your cell after sending the dinero so you did not have to hunt for that in prior texts or emails.

I just sent it. Thank you. It went to the account we had used before in that name. You might want to delete your email so that bots do not grab it for spammers. I put it back together with the Optimus block and KPX so it would be less effort to clean up when the SC DD block arrives. The temperatures are clearly higher, but not considering it was bare die with liquid metal before. Compare this (IHS and KPX) To this (EKWB direct die) running the same AIDA64 stress test for 2 minutes. Should have been at least 10-15°C diff. Basically the same result, within a margin of error.

That is very kind and generous of you. Thank you. Let me know how much to send your way for the shipping cost. You can send it USPS Priority in a small flatrate box. I really appreciate it. Your 12th Gen kit cold plate was made totally different than the first version I purchased that did not work well. I was impressed with the results you saw with the revised cold plate. It worked fantastic, just as the 10th Gen version I had did. I have started an RMA with EKWB and linked the thread you provided on OC.net and provided these two images with the RMA request to EKWB. Hopefully, they will do the right thing and give me my money back. I really like how it is made, but I do not like the poor results. They need to do something different. I left a comment on their YouTube video. I think delidding would help a whole lot more than you are thinking it will. As @tps3443has pointed out, sanding down the lip on the underside of the IHS so that the IHS contact surface mates with the die instead of the PCB is crucial. If the IHS drags on the PCB instead of swiveling freely on the die, the results will not be as good. If you do that and use liquid metal on both sides of the IHS even your sustained load temps will improve measurably. The benefit is not limited to short periods of stress. The stock solder only works well enough with near stock clock settings and an undervolt. Any kind of overclock with increased voltage causes unacceptable thermals with the stock solder.

First place to start is the delid. You're going to run hotter than you would after a delid no matter what you do. You'll never have excellent temperatures with stock solder unless you get freaky-lucky and get a CPU like one of those that @tps3443stumbled into that can run 6.0GHz with something like 1.250V. Once you have the IHS off you can easily lap it, or get an aftermarket copper or nickel IHS that is already flat.

Horrible, like all GPUs made before NVIDIA discovered a proprietary way to implement ray tracing as a means of extortion. Slide show... literally. https://www.3dmark.com/3dm/90334138 Please post a link to that @tps3443 I am not happy with my results. I am going to be asking EKWB to accept an RMA and refund my money for selling me something that functions the same as a normal water block on an IHS. My temps are not worse. They are a tiny bit better, but not direct die results like I would expect. Should be a LOT better, not a little bit. Temps with the Velocity2 block are almost the same as Velocity2 direct die block. Not terrible, but why bother with direct die if it is not impressively better? Makes no sense to me. Now, that said, I love how the EKWB part is made though. Very high precision and MUCH MORE pleasant to work with. I truly despise how the Supercool Direct Die block is made. Having to take it apart to separate it from the ILM is super retarded. They should come up with something less inconvenient and messy, because that sucks. A frame like EKWB uses would be a bazillion times better. It could be that. It is EXTREMELY heavy. It may have so much metal that it is normalizing temperatures rather than lowering them. White computer parts almost always strike me that way, even when it is crisp and clean looking. White is just not the right color to use from my perspective. It always sucks, even when it is not used in a disgusting way. The rainbow puke colors being added to the white simply create the impression of being a circus clown joke PC for children, and the rainbow merely exacerbates the magnitude of the ugliness. If you add stupid stuff like anime or cartoon images into the mix it then it becomes an object of contempt that is worthy of rejection and public ridicule. There has to be something abnormal that is being overlooked. What are you using for a cooler? Maybe that is the problem. Are you using the Intel ILM or a CPU frame? I am wondering if the CPU shape and the shape of the base of your cooler are going opposite directions and need to be lapped?

Yup, that was it. https://code.obsidian-pc.com/clevo-software/fan-control/ Weird. That used to work fine for me. Not sure why it's not working right. Way too long ago for me to offer any suggestions. (Obviously could not even remember their name now, LOL.)

There is a fan control utility from the Clevo place in the Phillipines (can't remember their name anymore, but I think Brother @electrosoftmentioned it once earlier in this thread). The guy's name was John, but I can't remember the business name.

Also no more yellow and red lights in Limit Reasons with Dell firmware cancer disabled with EFI variables overridden.

Yup, looks like a small improvement. I also disabled switchable graphics in the BIOS and forcing only use of the Quadro for graphics and for routing display output. Intel Graphics seems disabled now. HWiNFO shows zeros for everything. CPUx1: https://browser.geekbench.com/v6/cpu/413045 | https://hwbot.org/submission/5218488_ CPUx4: https://browser.geekbench.com/v6/cpu/413045 | https://hwbot.org/submission/5218489_

Mine is in excellent condition and I paid 1/10th (just under $500) of the orignal cost. I used an EFI shell earlier today and wrote new values to the firmware to unlock the CPU controls. I will need to re-run the GB6 tests to see if it was of any value.

Thanks. I dd not realize there was an updated version. It's not even that recent. I grabbed it. https://github.com/Orbmu2k/nvidiaProfileInspector

I think it is rare to find any notebook that functions correctly. Some malfunction worse than others, but they're all junk now. Some people just pay extra for their broken junk and hope that it is less broken.

But yours does have twice as many cores/threads and higher clocks than the older 7920HQ CPU, so that's good. They are both gimped 45W slopbook CPUs. What about your graphics compute scores for the Intel and NVIDIA graphics?

Feast your eyes on these poopalicious GB6 scores with my Precision 7720 turdbook. CPUx1: https://browser.geekbench.com/v6/cpu/409416 | https://hwbot.org/submission/5218152_ CPUx4: https://browser.geekbench.com/v6/cpu/409416 | https://hwbot.org/submission/5218151_ NVGPU: https://browser.geekbench.com/v6/compute/155841 | https://hwbot.org/submission/5218157_ iGFX : https://browser.geekbench.com/v6/compute/155857 | https://hwbot.org/submission/5218160_ #1 of 1 submission world records... 😆🤣

Although there is a degree of truth to this, I do not think it is accurate with computers, and ESPECIALLY not turdbooks. I think the revision should says something like this: If you pay extra, you are less likely to get screwed as badly as if you paid less. But, even then it's not totally accurate. For example, you'll pay more for an Alienware and get screwed worse than if you paid less for something else. But, yes Brother @PapusanRon White is right though... "stupid is forever" LOL.

That's why disarming us and having open borders make such "good sense" to those on the lunatic fringe. They want that. So, cool... what you're telling us is that you are normal then. 😉 On the other topic, I've never been particularly scared of creepy clowns, but they are a strange thing that just makes you feel weird, like something is really messed up. I'm not exactly sure why. I think that playing the part of one is a great way to get shot, stabbed, beaten up, or run over by a car, because some people get totally freaked out and lose their minds when they see one. Even seeing a "normal" clown freaks some people out really bad. I've even seen people close their eyes or look away when they see a creepy clown on TV because they get weirded out. I actually Googled it and it has been classified as a real condition. https://www.webmd.com/anxiety-panic/what-is-coulrophobia

Yup. That seems like not long ago but it actually was longer than it seems. It was 5 years ago and they were 1080 Ti. I know this because my only 1080 Ti SLI benchmarks were run that day. https://hwbot.org/submission/3839002_mr._fox_3dmark11___performance_2x_geforce_gtx_1080_ti_46904_marks It will be fun to see what I can do with it. I haven't used those tools before. They did not exist as far as I know the last time I owned a Red GPU, so I will need to grab them and figure it out. Thanks, Brother @Rage Set.

Great news, Brother @Papusan! Looky here... you can get bling bling for your crocs now! Hurry, before they are all gone. https://www.amazon.com/dp/B0BWTHRRW4