Mr. Fox

I never had any "problems" associated with memory temperatures with the 3090, but I never tried using it in a SFF hotbox that suffocated it. I suspect the 12VHPWR connection was never a problem (I do not remember ever hearing about one melting on a 3090 Ti) because everyone basically had the ugly stock pigtail as the only solution. The 40-series brought custom cables and a relevant aftermarket parts presence because it wasn't something applicable to only one overpriced and arguably overrated GPU model with very limited market share. The tacky-looking stock 12VHPWR pigtail has no flexibility issues. It's just a whole lot uglier than anything else available. And, that won't matter in a chassis where nobody can see how ugly it looks. What I find most sad about the stubby eyesore stock adapter is nobody learned anything from the 3090 Ti, or they just didn't give a damn, because all they needed to do to make it less of an aesthetic atrocity was add another 8 to 10 inches to the length of the cable to facilitate a cleaner-looking build. If you could hide the nasty-looking "knots" where the stock 8-pin PSU connectors attach to it, the aesthetic value of the aftermarket cables would be dramatically diminished. I would say minimal difference unless you compare against something measurably better like a FTW3 or KPE or Galax HOF. It may be more important to focus on what fits in the limited space and how the cooler design works best with where you are going to install it. In a SFF scenario, the FE might be best suited to exhausting hot air through the I/O plate like the FE does, rather than blasting it sideways around the perimeter of the heatsink into the chassis like most other coolers. As best I can tell from some limited-interest Google searching, there are no mini-ITX motherboards with respectable overclocking capabilities. They are 2-DIMM configuration (desirable) due to size, but they all seem to have limited overclocking potential and wimpy 10-phase power delivery designs. The focus seems to be on size reduction rather than performance.

Yes... more powerful, more VRAM, easier to resell and holds its value better. Might not "need" an upgrade to 4090 if the 3090 ticks all of your boxes. I'm not sure. Not that I am aware of, but I have never paid any attention to 3080/3080 Ti options.

I would recommend going with a good 3090. You might find you have no compelling reason or need to upgrade. It will also be easier to sell if you decide to sell it later. I would recommend not going with a 3090 Ti because they use the 12VHPWR connector. See in-line comments inside of quote

I think some people might be getting mixed messaging about the black screen issues that some of NVIDIA's newest drivers and DWM.exe are causing and thinking their cable is burning. Some of the latest and greatest DCH filth drivers can cause users to experience random black screen issues.

Maybe it is only an odd coincidence... Don't know since this is the only 13900KS sample I have seen, but it seemed like the rubberized adhesive and solder were much thinner and a lot easier to remove on this one. The delid itself was easier (required what seemed like less force) and the cleaning and prep seemed easier.

Electronics and technology are a racket and organized crime ring.

My bad. The E-cores were not at 48x. Now they are. The extra 200MHz on the baby cores does make a difference. Same voltage (1.400V vCore).

Probably planned that way. When the Green Goblin still supported SLI it took 3 or 4 generations of improvement for a single GPU in the same performance level to match 2x SLI. Money didn't come as easy for them as it does now. They actually had to earn it by releasing something special enough to beat the SLI setup. OK, so 60x all P-core, bumped ring to 51x. vcore needed is 1.400V.

Looks like vcore of 1.290V is the bottom for 58x on all P-cores. Quite an improvement from the 1.350V needed for 56x on the 13900K. It might go a little lower with adaptive voltage, but probably not worth burning any more calories on it. I think this is good enough. Time to go up on the multis.

OK, so I just delidded the 13900KS. Much better temperatures. Still testing, but 58x on P-cores is fine at 1.300V static vcore. Will see if I can go lower, then see how much it needs for 60x for all P-cores.

Or, depending on how you look at it... We can ask the same question of NVIDIA and AMD now... $300? Really?

Intel Arc A770 is the better GPU here. The RX 7600 is cheap GPU for people that need basic output and the ability to play YouTube videos, or smartphone games on PC such as Tetris, Flow Free, Bejeweled, Angry Birds and Candy Crush.

I did not list the additional items. Only the 3090 Kingpin was listed there, and it sold almost immediately. https://www.overclock.net/threads/rtx-3090-k-ngp-n-hydro-copper-block.1805258/page-2#post-29170391 I did not list the NUC there for a couple of reasons. (1) doubted the people there would be interested in a NUC, but may not be an accurate assumption; and (2) @ryanhas expressed interest in it. He has been cash-strapped and selling it right away is not necessary for me, so I did not want to list it a bunch of places and have it sold out from under him. That being said, if someone else in this community wants it, because it is listed here it is fair game. He doesn't have exclusive dibs on it here. If he or someone else here doesn't grab it in the next week I will list it on OC.net, Mercari and eBay. Selling it here or on OC.net not only gives people a nice opportunity to get something at a great price, but also affords me the opportunity to avoid getting financially raped by Mercari or eBay and PayPal. When I list something on eBay or Mercari, I generally ask 15-20% more than I actually want to offset their punitive fee structure. That gives the buyer a bit of a cushion to "chisel" me down on price without me taking it in the shorts. @Rage Set might be interested in the Strix Z690-E mobo and 13900KF for a custom build. If his customer(s) are not interested, then I will post them here (combo sale) once that is determined. Like the NUC thing, my goal is always to show preference to friends first, community second, and strangers last whenever possible. I will likely throw in a 32GB Hynix M-die DDR5 freebie if the buyer is interested in that combo.

The headphones and docking station are gone. The NUC is still available unless @ryanis still planning to grab it on the 28th.

That looks really great, bro. Super clean. I'm not seeing the mess that you referred to LOL. Yup, trying to minimize dust was the reason I left the open bench and went back to an enclosed case. Had it actually worked out the way I expected, that would have been great. In reality, it didn't solve my dust problem at all. It might have even made it a little bit worse. It's hard to quantify that since there's no way to measure the dust. Ultimately, I ended up having more surfaces to clean that were harder to get clean due to having more nooks and crannies to get dirty. Everything on the outside gets dusty. Everything on the inside gets dusty. All of the surfaces seem to need an equal amount of manual cleaning.

The MSI Suprim is the same. The sensor pins are next to the PCB. Definitely going to be a nice chip once I delid and go to bare die. Using 1.341V static vCore at 58x on P, 46x on E, 50x on Cache, 8200 on Memory. The 13900K needed 1.350V for 56x on P-cores. It is insanely hot running without a delid and KPX paste on the IHS. Gonna fix that problem next.

Jay had some kind of surgery, so Phil did the video. Apparently, Jay is going to be out of commission for a bit while he recovers. No idea what the deal is, but he looks medicated in that video. Here is a comment that explained the situation. While buying an 8GB video card in 2023 is not very smart, and it was not smart for NVIDIA to offer one, I do not agree with the logic that buying a second hand AMD card for slightly less that offers similar performance is an equally good or better option. I think it is a bad option and I would buy a 4060 Ti before I would spend a similar or slightly lesser amount for an obsolete 6700XT / 6750XT. All things being mostly equal, I will always choose an NVIDIA GPU over a similarly priced AMD card if for no other reasons than better drivers and better features, and fewer bugs.

We need these emojis added to the button list in the lower right corner 💩 💢 💯 I would use the pile of poop for that one, LOL

TL;DR? ^^^ What he said, LOL... ^^^

Definitely looks amazing. I was (and still am) torn between it looking incredible and losing the huge amount of space it consumes. If you remember when you offered it to me a couple of years ago, I got out my tape measure and started trying to figure out how it would work and decided it was too big to fit. I was thinking with my head instead of my heart and being pragmatic at that point. The reality is it will not fit anywhere comfortably, but I am still seriously considering moving the guts from my work computer into it and getting rid of the 5000D Airflow. It also looks amazing but it's too small. It's like the exact opposite problem LOL. The only thing that gives me pause is the fact that if I need to work on it I can easily, with one hand tied behind my back, spin the 5000D around or lay it on its back on top of my desk and it's no problem. I don't have to move other things out of the way or clear space to move it. Moving the DG-86 is definitely not something that can be done in a nonchalant manner. It requires strength, both hands, and a certain degree of planning (and moving other things), LOL. The empty chassis weighs more than the fully assembled 5000D. (I am only assuming this, but I suspect the gargantuan size and weight is why you are not using it instead of the Lian Li case you are using now.) The DG-86 is too large to fit under my desk. It's about 6 inches wider than my desk is deep. This is due to the rear IO chamber, which is also one of the features I like so much. So it's like a two edged sword in that respect. I wouldn't want it under my desk even if it would fit, because that is just way too inconvenient and inaccessible. And what's the point in hiding something so gorgeous under a desk where nobody can appreciate its beauty? That's kind of stupid. I have always absolutely hated the idea of putting a PC under my desk or in an enclosed compartment inside of the desk. It's so large that having it on my desk leaves me with no workspace, which is already at a premium. Going from the Level 20 XT to the DG-86 was jumping from the frying pan to the fire. The footprint is the same, but the DG-86 is a whole bunch taller and much heavier. The added height isn't as difficult to deal with as the amount of space the footprint consumes and the complexity involved in moving it. The one major advantage to the Level 20 XT is its 360° access, horizontal motherboard tray, it becomes the functional equivalent of an open bench with the glass side and top removed that is just as massive as the DG-86 in terms of desk footprint. What I really need is a larger office but that is never going to happen. If I had a larger office space I would probably do several things that I can't do now, including using both the DG-86 and Level 20 XT simultaneously and buying a second chiller or phase change system. Using external pumps and radiators contributes to the space problem, but I am not interested in changing that. The external pumps and radiators approach is the only way to fly, IMHO. The benching rig is still definitely staying open air bench like it should have stayed all along. Functionality trumps aesthetics in that respect. The space the open bench takes up is only 15x15 inches. It is so much more compact that I've even considered the idea of getting another open bench and not having any enclosed cases. The unrivaled convenience of using an open bench is absolutely compelling. Moving it is almost as easy as moving a milk crate. Just grab it by the side handles and move it (after disconnecting the wires and two QDC fittings, of course). Having used few enclosed cases, all of which have been stunningly gorgeous models, has made me more appreciative of the convenience of an open bench and less influenced by something because it looks nice. If I were a gamer that valued form over function and built something, set a modest overclock, then put the side panels on, left it alone and just used it, that would be different scenario. I would be using hard tubing and doing things that gamers do that valued style over substance. But, I am not a gamer and I can't put something together and just leave it alone. I am uncontrollably compelled by my mental illness as an enthusiast to mess with it. Constantly.

Thanks. My GPU is over those thresholds at idle. It is like 12.900V. That might be the 1000W vBIOS I am using though. Not sure if that affects input voltage or not. I will check with the stock vBIOS and see if it changes. So, I got the 13900KS from Germany. Exactly as advertised. It is in the buggy Z690 Apex right now. Needs a delid real bad, so I have it capped at 60 on all P-cores. Will do the delid and move it to the Dark mobo in a day or two. Looks like a strong CPU...

Hmm, I have not seen that. Which sensor is it in HWiNFO64?

What we do not know is how many have been sold versus suspected of failure. If CableMods sells volume of 10:1 or 20:1 compared to the other brands and 1 in 20,000 sold by CableMods results in a problem, it doesn't mean the other brands are better or that CableMods is worse. It just means we don't have enough data to speak with authority on the matter and have to resort to speculation. (I am just pulling numbers from the air for discussion. I don't think anyone has any real numbers.) That doesn't mean that anyone should think differently about one brand versus another, or feel more comfortable using FASGEAR or MODDIY instead of CableMods. The only thing we can be certain about is that we don't know what we don't know, LOL. As painful as it was to lose them in the GPU race, I think it might have been a smart thing for EVGA to do at this point. Putting myself in their shoes, I might have made the same decision. Other than it being a favor and an act of sacrificial charity, just because they want their customers to have better products available to them, there would be no true business reason for them to continue in a situation where the conditions were poor and profits too low. It is dangerous for consumers that EVGA might not be the last one to bow out and excuse themselves from the nonsense if things continue on their current path. There are a couple of things that don't make good sense to me relating to the 12VHPWR connector. The sensor wire is very thin and delicate. Why? Is there a reason it is not the same guage as the rest of them? Leaving it free-floating places it at risk to some degree for getting broken or cut. But, confining it also makes is susceptible to kinking or breaking because it can't lengthen or shorten itself at the same rate as the heavier wires when the cable bundle changes from being straight to having bends in it. (That is one advantage to the flat cables for sure.) The 12VHPWR connector itself is dainty and fragile. Why? I can't identify a legitimate reason it needs to be. They could make it 50-75% larger and stronger without losing anything, and it would still be smaller than three 8-pin connectors if a silly goal to make it smaller in size is what is driving the idiocy. (Which would fly in the face of logic looking at the rest of the GPU.) The wimpy design of the 12VHPWR connector seems especially out of place on graphics cards that are some of the most robust, high-powered and overbuilt GPUs the world has ever seen, yet they put a wussy baby-girl power connector on them. That's real stupid.

I have not seen any issues with my cables, end terminals or GPU socket either. All still look like new. I frequently check mine for heat by touch and using an infrared thermometer, and I have never seen anything unusual so far. All of the hyped crap on YouTube and tech web sites makes many people feel less than 100% confident, since we can only speculate on what is causing it. Some of the speculation is probably accurate based on logic, but without evidence. I really don't think it is happening because people suddenly forgot how to plug in a cable securely, especially after the first round of negative publicity. I suspect most people are cautious to a fault and still not feeling any warm fuzzies about it.

I have not seen one, but a cable with the terminal on the GPU end having the same variety of 90° orientations as the Cablemod adapter would be good. Having the cable connect to an adapter that connects to the GPU provides a potential additional point of heat-generating resistance, and in the worst case scenario, increasing the opportunity for failure/damage. Even the crappy stubby adapters that ship with new GPUs provide an additional point for possible failure. Those adapters would not be as undesirable if they had cables that were 12 inches long. Their stubby design is what makes them so stinking undesirable. They produce an eye sore in an otherwise clean-looking build. Maybe they made them ugly on purpose so people would want to spend money on new ATX 3.0 PSUs, better aftermarket adapters or 12VHPWR cables. (Yes, that would be collusion and a hybrid scalping technique to spread the money around to everyone that wants to have their hands in the Ada Lovelace profit jar.) What I meant was loose in the middle, but not disconnected at either end. I am not sure what anyone else meant if not that.