electrosoft

3090 is the best bet if only for the 24GB. Dan C4 works with a lot of GPUs but really shines with Founders Editions so you could target a 3090 Founders. It can even accommodate 4000 series. 13700k is the best way to get a better 12900k (which you are already using) 2-slot Z790 itx = much better DDR5 performance. You will most likely be able to hit 7200+ depending on sticks.

😁🤣 I don't know what Nvidia and to a lesser extent AMD were expecting.... For the average gamer who at least even somewhat keeps up on market conditions and does a smidge of comparative performance analysis, they know this is garbage offerings from both companies. Nvidia just smells a bit more. AMD at least changed course right at launch and lowered the price of the RX 7600. 7900xt and 7900xtx have all come down in price plus a free TLOU game code (even the RX 7600!). Some 4080's are down to 1099 AND free Diablo IV ($70 value). Free game offer extends to the 4070 but not the 4060ti (boo!) Of course where able (IE unverified purchasers) it is getting review bombed everywhere.

"And the understatement of the year award goes to....well will you look at that.... @Mr. Fox!" Seriously though, that is a much better specimen than your other chip. Remember when the techtubers and sites were commenting on it was odd for Nvidia to introduce the 4090 first before the 4080 as that was out of their normal process (xx80 first, then halo product)..... As for the 12VHPWR, the problem is on the GPU end but if every burnt connector can be verified to come from non ATX 3.0 power supplies there may be a connection. On the other hand, I suspect the sheer amount of non ATX 3.0 power supplies powering the 12VHPWR connector vastly out number those with the ATX 3.0.

I mentioned this a few days ago in my post along with here too showing strategically limiting their memory, memory bus and core counts that Nvidia (and AMD to a degree too) is forcing their GPUs in their product stack into tiers along with price hikes. ADA is too powerful and if they were to offer a 4060ti series die with a 256-bit bus and even 12GB (let alone 16GB) at $399.99 with the same shaders as the 3060ti you suddenly do not need much else to game at 1440p for many use cases. How do you create a center stack before up-selling the 4080 and 4090 for 4k then? Everyone thinks Nvidia should have made a 4060ti with 16GB and a memory bus and core counts like the 3060ti for the same price but that just makes no sense business wise except to the consumer who always wants more for less. Sub $500, 1080p (With varying degrees of setting enabled depending on Vram/GPU aka 4060ti 16GB at 499.99) Sub $1000, 1440p (" ") >=1000, 4k (" ") "More for Less" is not something Jensen will ever subscribe to but with poor economical conditions world wide, GPU saturation and a slowly turning tide in resentment on GPU pricing overall, hopefully a lesson can be taught.

Let me pre-empt this by saying I use a Torque screwdriver for establishing optimal tolerances in many scenarios where torque is variable and not preset to establish precision tightening. This includes primarily laptops but also desktops along with other tasks when needed. This approach is also how I documented and found out it was the uneven pressure of the lower left CPU retention screw on the X170SM-G that was problematic in comparison to the others causing post and crashing issues. With that being said, plenty of users hand tighten and working within that dynamic margin of error have manage to drop their temps. Others, like Jay, did not really experience a worthwhile change. When problems have presented themselves (usually no post, crash during thermal expansion and faulty memory channels), they go in and manually loosen/tighten till it works and then it works for them (aka well enough). Some contact frames do come with measurements/recommendations. There are also many trial and error guides on line too. There are two trains of thought here: The idea of a contact frame is a waste and not needed even if mounted properly. The human error involved in hand tightening is more problems than it is worth. What is your opinion on both?

Rando repair video and Greg noting the 12VHPWR cable was not fully inserted from the owner....

Ah ok. I was just checking if there was some other unknown issue outside of human error/torque. Some see no results while others do with the contact frame. As always, YMMV. I always liked Jay. Just a dude who likes cars and tech and made a channel about it that took off. He never takes himself too seriously and it looks like they have more fun than not making videos.

What issues did you find with the contact frame outside of human error in regards to torque?

Yeah he definitely was out of it. I do not agree with all the backlash to a degree. I think the 4060ti is a fine card just not at that price point for the performance uplift generation to generation. Seeing as Nvidia already slowed/shut down 4070 production already I suspect they may feel the same wrath with the 4060ti. With VRAM being such a central issue (feeling yet again vindicated from late 2019/early 2020 when I was getting jumped on about VRAM issues and problems even with general allocations being high), an AMD card makes sense especially a 6750xt at 319.99 with 12GB vs a $399.99 4060ti with 8GB: A 16GB 4060ti at $399.99 and 8GB 4060ti at $349.99 (along with 4060 at $299.99) would have reviewers very happy at least due to the memory bump even if the overall performance vs the 3060ti is rather poor.

Nice! But seriously.... I say this with the utmost respect.... GO GET A 4090 ALREADY! 😁😆 No surprises here. Nvidia is re-aligning everything for the future and pricing since they got a taste of what consumers can and will pay during the mining pandemic times and to properly align their cards in their product stack you shift prices upward, cut dies, bandwidth and memory. You are no longer guaranteed a healthy performance bump along each tier even if Nvidia has to neuter their cards to give it to you. You won't get a *worse* performance bump tier vs tier (3060ti vs 4060ti) but it won't be the generational uplift we're used to in the mid to low tier anymore....and you'll pay for that privilege. 4060/ti is meant to be a full blown higher refresh 1080p card and that is what you'll get...and like it. Want Ultra settings too? Well, we have a 16GB model for that coming....aren't you happy? What? You want to do 1440p? Well, we have the 4070 and 4070ti....more memory? Fine, FINE....here is our $1200 4080...you are so lucky! Looking at HUB's review, on average it is ~8.8% faster at 1080p and ~5.4% faster at 1440p vs the 3060ti. It is a high refresh enabled card at 1080p and playable at 1440p. That is pathetic generation to generation wise. Nvidia is Intel ~5yrs ago but with the technical advancements to back it up but consumers don't forget and the mindspace thoughts about Nvidia are slowly shifting. The moment they falter and AMD (or Intel?) offers at least equally compelling products they will feel it just like Intel is feeling it now.

This build hurts my head.... $100k budget for a lot of BS but at it's core they select an EVGA Z790 Kingpin, 13900KS and G.Skill 6000 memory (?) 2x32GB and it won't run at XMP (or has issues) so they drop down to 5600..... That much budget I am definitely: Buying a binned out the wazzo 13900ks Drop down to 2x 24GB SR sticks / Splurging for some 8000 sticks Delidding the 13900ks Tuning the snot out of the memory, CPU and GPU (in that order) with all that cooling Might as well water cool the memory too... Heck with that budget I'd probably fish around for a killer 4090 sample too

I think the problem with the free floating design (potentially) is the floating leads and potential movement in the docking like @johnksss alluded to previously (as if he is rarely or ever wrong 🙂 ) In hindsight, I would have tested my original moving the wires manually to see if it would work and/or cause a shut down. This time during my next system overhaul I'm going to purposefully play with the lines/pins connected under a few different load levels and see if anything happens. There is zero movement in the MSI adapter (aka Quadro Force Tentacles **ATTACK**) Is there an all in one 12VHPWR 90 degree adapter and line on the market?

Yep and they do move slightly back and forth tugging on the cable. Replacement showed up today and visually it looks identical and the sense pins actually move a tad bit more and the lines are still free floating.

Re: Buying positive reviews Ugh, Asus can't keep shooting themselves in the foot....sheesh.

@johnksss @Papusan the sense cable makes more sense. While it was technically connected, the cabling was floating on its own versus everything else. Here is a pic of my actual failed Cablemod cable (before I cut it in half as per their instructions). This is the EVGA specific cable and you can also see how they just attached leads based upon which PSU you used:

My new Cablemod replacement is still en route direct from China. My January purchased cable when I finally got around to using it last month shut down my entire system under any significant gaming load (>=300w). Stock MSI cable continues to work as advertised even >=530w. This was a specific EVGA cable for my P2 1600w (and compatible with my P2 850). I held onto the old cable (albeit they had me cut it in half and send proof of destruction before shipping a new cable) so I'll be able to physically compare the new one with the old one. Re: Asus - If they had just worded their Beta BIOS information better most of the heat they're catching would have been eliminated.

Microcenter now sells Nvidia Founders cards too? https://www.microcenter.com/product/661136/nvidia-nvidia-geforce-rtx-4090-founders-edition-dual-fan-24gb-gddr6x-pcie-40-graphics-card Looks like Nvidia is also selling them directly now again: https://store.nvidia.com/en-us/geforce/store/?page=1&limit=9&locale=en-us And as always, Best Buy: https://www.bestbuy.com/site/nvidia-geforce-rtx-4090-24gb-gddr6x-graphics-card-titanium-and-black/6521430.p?skuId=6521430 They always seem to still be out of stock more often than not though.

I've been researching the Caselabs "luggable" too. I do like the 24" 144hz 1080p display but the idea of it recessed so much during use by design with all the components right in front of you might be a heat bath to the face while using and an odd viewing angle. In the video they prop up the back (side?) but then the display is really getting up their off the table for viewing angles. I wish they had a more secure GPU mounting system with an adjustable securing mechanism in the front of some sort. I do like that it is sealed air tight when closed (they douse the whole thing with water for show) and it is clearly geared for travel. Not sure it would survive being checked luggage at an airport or train travel though. Keyboard and mouse have to be stored/shipped elsewhere. Pros really are you can cram just about anything in there. Looking at it trying to visualize how my MSI Liquid Suprim 4090 would work in there with the AIO. Maybe strap it to the inside for travel then when gaming just lay the AIO on the table. Hmmm, that would work.

Northridgefix and the "influx" of 4090s which were actually directly from Cablemod accumulated over a few months from customers unable to get OEM repairs? That was a little bit of egg on his face....but he's so direct it's refreshing. 🙂

Continue production and adjust ASP down or slow/halt production and keep the price the same. If the end result is the same gross revenue but higher profits for Nvidia, it makes sense versus having a minor version of the 3000 glut which proved to Nvidia they could commit to the long con...er....strategy and eventually move those units at inflated prices or at worst MSRP 2-3 yrs removed from launch. WIth Nvidia and AMD to a degree, the problem is these newer generation cards at lower tiers are too powerful to lock them into "1080p" mentality. They are quite capable of even at 1440p but that isn't their target market. If you want a 1440p card, you should be stepping up to a 7900xt / 4070ti type class card according to AMD and Nvidia. How do you add clarity to your stack segmentation and force consumers to spend more to get a proper 1440p and 4k experience? Limit the VRAM and cap the bandwidth.... 8GB = 1080p low/medium, 1440p low 12GB = 1080p High/ultra / 1440p low/mid 16GB = 1080p Ultra, 1440p high/ultra, 4k mid/high 24GB = 1080p Ultra, 1440p Ultra, 4k high/Ultra Then add insult to injury by pricing them up even more giving you less bang per buck vs last generation. I could never stomach having a SFF as my main desktop. I spend way to much time in there testing components and swapping things around...plus I love having something beefy on my desk. But the more I think about it I DO like the idea of being able to extract the two most expensive components (or all of them if I wanted to go to that extreme), transplant them into a SFF which has the MB/PSU already installed and ready to go with riser slot in place and swap over my CPU and GPU and take it with me and get the same level of power. If memory was more costly, I would swap that too but luckily with these used A-die 2x32GB I picked up and tuning (which are shaping up to be better than the M-die sticks), I have these extra M-die sticks ready to rock too. I just need (~550-650) SFF case w/ Riser SFX PSU ITX MB And when heading out to travel, I'd swap in (in my desktop or items I already have on hand): 4090 GPU 7800X3D CPU Memory 1TB M.2 Get back from travel (work or vacation) and move the components back in to my main rig...BAM....back up and running.

Well I do game quite a bit when on the road and I prefer a real keyboard and mouse vs the laptop variants. When traveling I (solo) or we (the wife and I) are stationary for days. I'm trying to strike a balance (hence the shared GPU/CPU between big desktop and sff on the road) minimizing costs. 15.6" is too small and I would end up taking a 20-24" display with me anyhow or buy a Alienware m18/ Acer Predator 18 level laptop. When I game when traveling I like a no compromises experience just like at home so I would end up picking up at least a 13980hx (or AMD equivalent) and 4080 at the minimum. So factor that cost vs having a compact sff frame to take instead and sharing a desktop CPU and GPU in it. When it's time to travel just swap the components over to an already set up chassis with an itx MB, memory (32GB vs 64GB), PSU that just needs the GPU and CPU swapped back in and off you go. Another option was that luggable option using the same approach linked earler this week.

Instead of investing 2500-3000 in another high powered, limited laptop this is what I'm thinking of building out for travel (something along these lines) with a 4090 FE + 7800X3D. https://formdworks.com/collections/t1-1/products/t1-sandwich-kit-titanium-color?variant=42966874226935 I would just use air cooling on the 7800X3D and ideally I'd like to be able to swap in a desktop 4090 FE (and potentially CPU) for travel then swap the CPU and GPU back to desktop when I get home. Wireless gaming KB/M along with a decent and most portable display. If I went this route, I'd sell my over achieving Suprim as there is no way I'd keep 2x 4090s. I would just use one and when I do travel I'd pluck my 7800X3D and 4090 FE and pop them in and off I go for what amounts to maybe an additional $500-600 in costs total for the sff, PSU and itx motherboard. I would still keep my NH55 for true portable use when needed but on travel with the wife let her use it for WoW. Comments? Suggestions? What would you do?

A prototype was seen months ago. It is massive and the main selling point is the 48GB of memory. It is going to cost probably $3k and is aimed at professionals. ~11% uplift is about right between the 4090 and a Titan class card for GPU performance. Might even be a hair slower.

So true brother. 😞

When DTRs were the only option, it was what it was and if you needed a laptop your pool was limited but as laptops and notebooks got smaller the average consumer spoke up with the spending power, the bean counters responded and production shifted accordingly. The market for the type of DTRs we like here *is* a fringe segment and always has been. Even modern DTRs (Titan, m18,g18) are but a small segment of the overall market even modernized to a degree.