Mr. Fox

I have no idea. I haven't seen anyone overclocking 32GB dual rank DDR5 modules, so you might be the $300 guinea pig. I also haven't seen anyone overclocking a 4-stick DDR5 setup. Seem like all of the overclocking, at least that I have noticed, is being done with single rank 16GB modules in a two-stick system.

I agree. The SK Hynix A-die is way too expensive right now, and the degree to which it would be beneficial is questionable. @tps3443 I would say wait a while. I could not tell any difference in performance with SK Hynix M-die at 6800 versus the Samsung B-die you purchased with the mobo running at 6400. At 6400 and equivalent timings the M-die Hynix and Samsung B-die perform within a margin of error of the same. The only way I could measure a difference at 6800 was with a memory benchmark. I ran some game benchmarks, Cinebench and 3DMark and the difference was not really detectable. In most things the performance of the DDR4-4000 on my Strix D4 matches the performance of my DDR5-6800. I'd say when the A-die price falls to about what we are paying now for M-die it will be worth it at that point. If you are just wanting it for impressive memory benchmarks and don't mind paying a ridiculous amount for the A-die, then by all means do it just for shiggles. I want it for no reason other than the AIDA64 benchmark, but not bad enough that I would be dropping a bunch of money on it for the reasons stated above. The overpriced A-die is a very poor value and there is no bang for the buck at this point.

The RTX 4090 is far too expensive even if it does offer double the performance of a 3090. This is abusive Kingpin/HOF overclocking enthusiast niche GPU pricing on belly-button video cards with cancerous firmware made for Joe Blow the gamerboy. They will have gimped performance. And, if they pull more shenanigans like they did last time, there will be a 4090 Ti and you will be stuck with a second fiddle graphics card that you paid more for than it was worth. Even if it were priced appropriately, it would be prudent to wait and see what cards are going to have good water block support and which ones will not. An enthusiast version of any GPU is going to be boost clock impaired and thermally gimped if it relies on air cooling or a hybrid setup. My confidence in NVIDIA or AIB partners to do the right thing and not deliberately rape people willing to cough up the cash is near zero percent.

But, I am actually all for this, LoL. Seriously, I am for real. I am opposed to normalization and forced equality in any form it takes. If there is no opportunity to establish superiority, rise above the herd, be remarkably better than average, take a dominant position and send the gamerkidz away crying because they want to and don't know how, then I have no interest in it. The joy of overclocking is lost no matter how good the "normal" performance looks when everyone gets roughly the same experience. It might be good for ordinary consumers, but it is no longer exciting to own for an overclocker because it becomes ordinary. Using chilled water with the 5950X provided little benefit because the silicon was already close to maxed out, which made owning it very boring and unstimulating to me as an overclocker. It irked the snot out of me that I could not get it past 4.9GHz stable on all 16 cores/32 threads on chilled water even though the 7980XE was stable at 5.3GHz. That made it feel inferior even though its clear improvement in IPC gave it an edge in benchmarks over the much older CPU.

There were a lot of things I despised about the 5950X, but PBO was enough that had it been the only thing wrong with it I still would want nothing to do with it. It sucked real bad. I know the lazy gamer kids that don't want to learn how to overclock the right way like it, but it makes no difference to me what the everyday user/gamer likes. It's AMD's job to care about that. That they do like it sends a strong message about it. I hate Intel TVB just as much, and for the same reasons. I overclocked it manually and got much better results, but some of the fanboyz seemed irritated that my results destroyed their hypothesis that PBO mediocrity was better.

Price: $250.00 - Sold for $275 on eBay Condition: Used for a few days, so like new Warranty: Lifetime through manufacturer / none from seller Reason for sale: This is a spare kit that I do not need. I am using generic modules with water cooling and don't want/need RGB Payment: PayPal, Zelle Item location: San Tan Valley, AZ Shipping: Lower 48 US - First Class Package for free, or $5 USD if you want the package to be insured International shipping: Actual cost plus any taxes, duties, VAT, etc. I am unable to estimate these costs. Handling time: Next Business Day (I am on a business trip until this weekend) Feedback: eBay Profile; OC.net MarketPlace Specification: TeamGroup T-Force Delta RGB 32GB (2 x 16GB) DDR5 6200 (PC5 49600) These are built with SK Hynix M-die IC and are stable with tight timings up to 6800 MT/s with less than 1.500V. https://www.newegg.com/team-32gb-288-pin-ddr5-sdram/p/N82E16820331744

That makes sense. Had you been a normal everyday misguided "newer is always better" laptop muppet chasing the latest BIOS "update" garbage you would have taken the same hit as every other beguiled victim in the sheeple herd. There were also some BIOS mods on HWBOT with the CPU microcode removed completely. Everything worked fine for me on those BIOS mods without it. I also disable Intel ME in the EVGA BIOS. EVGA is the only BIOS I have identified with an option to disable Inel ME. I haven't found anything detrimental with the Intel ME disabled in the BIOS. Everything I can see functions normally when the Intel ME doesn't exist. We have been conditioned to believe both of those things are necessary or required and neither of them are. You can also delete the CPU microcode for Intel and AMD that is part of Windows and get updated through Windows Updates. It is likewise not needed. The intelligent approach to firmware updates is to ignore them unless it fixes something important that is broken without it. It is rare for that to be the case. I never identified enough of a benefit to Adjustable Bar support to qualify as important. In some circumstances it serves no useful purpose. The downside to emerging technology, like DDR5 architecture, is that improvements in firmware can be relevant to that element while being simultaneously detrimental because of the payload of performance-killing garbage bundled in the same firmware revision.

This happens when people that are OCD about "security" are allowed to determine the fate of those that have no reason to care, and those that are willing to live with vulnerability in favor of better performance.

I don't think there was a BIOS for 13th Gen when I sold it to you, but you can do the recovery flash procedure with the Unify-X with no CPU or RAM installed, same as EVGA and ASUS. You rename the file MSI.ROM and put it on a FAT or FAT32 formatted USB stick, then use the button on the rear I/O panel next to the designated BIOS flashing USB port. I suspect it would still boot and run without an issue with the 12th Gen BIOS, but it might not have full support or total functionality until you flash it. They are similar and the outcome can be the same. I believe the former suggests the demonstration of a lack of good judgment and the latter refers to the absence of capacity to exercise good judgment or discern the difference, so yes... it could be both, but I think more often than not their bad decisions are deliberate and they know what they are doing is messed up. They don't care.

Filth. Digital dung. The excuse given "to ensure better compatibility" is a blatant lie. What they meant to say was "to ensure we maintain the power to own you and your PC" because their wickedness and treachery is unlimited and far-reaching.. Totally a Dellish Nazi power play.

It is in the benchmark settings/options page and is a local machine setting. Turn off the option to automatically hide results. If you sign in at 3DMark.com your submissions should be stored under the My Results page. You can see all or filter by benchmark. If you go to a benchmark result online you will see buttons to hide/unhide and delete if you submitted it associated with your account rather than anonymously.

I scratched them off my list a long time ago. My first and last Chai Tea mobo cured me. I've blacklisted them in my mind. It's always possible that they can do something to earn their way off of that list, but right now I am not aware of a way. They deserve to be blacklisted if for no reason other than their warranty terms are unusually hostile toward customers.

He mentioned how worthless the ASROCK firmware is on the system available to him. Maybe the option is missing.

Thank you. Makes me happy to know that information that I share gets put to good use with the people I care about.

Here is this... https://extremehw.net/all-reviews/sabrent-rocket-ddr5-16gb-u-dimm-review/

I preordered from NewEgg as well... 13900K not KF. And, I got an email saying the product is delayed but my pre-order will remain in effect unless I choose to cancel it. 4090 is extremely unlikely. 4090 Ti is also extremely unlikely. An AMD GPU even far less likely. I don't really see much point in a GPU upgrade right now. Too much money for meaningless points on HWBOT. For now it remains to be seen. It will depend on my ability to exercise common sense and self-control. It may turn out that I demonstrate foolishness and stupidity once again.

ASUS definitely offer feature-packed options, but they have had a horrible track record of failures including a fire and killing a 100+ SP CPU during one of the failures. It is always with great fear and intrepidation I purchase anything from them. When everything goes welll they are outstanding. When you need their help, they suck. If it were not for their poor reliability and lousy warranty service they would be the best. But, most of their competitors offer equally lousy warranty service, so I can't single them out on that character flaw. What is most interesting about my experience is their flagship products are the only ones I have had reliability issues with. The Prime and Strix mobos have been very dependable, but less capable.

It's got to be exciting. Motherboard is a tough one. With EVGA no longer in the game for AMD none of the acceptable options have respectable warranty support. Not to mention you'd probably be waiting for a year for the first parts to drop. EVGA is always painfully late product releases. Probably should stick with ASUS as much as I truly hate to say that. Hopefully they'll have a good two slot DDR5 option like the Apex. Are you going to get a delid tool and bare die frame from der8auer?

That is how it should be. There is no place for being a gentleman in business. You do what is best for yourself, your company/employees and your customers, in that order. You do what is worst for your competition and do it deliberately.

Awesome. Congrats. Where did you get the A-die? Splave at HWBOT or somewhere else? I have been looking on eBay and not finding anything, and I am not willing to pay more than about $250 USD for a pair of 16GB sticks. What do the AIDA64 read/write/copy/latency numbers look like at 7200? I have found M-die to typically cap out around 7000 as well with higher voltage, 6800 with less than 1.500V.

That is still going to be a performance-limiting factor, just as it is for GPUs. Not being hot enough to kill the chip doesn't mean it is ideal. They can say that it is OK all they want to, and change the firmware to let it run that hot without throttling, but it will shorten the life of the silicon and will limit overclocking headroom as a matter of natural consequence. Taking the right steps to make it run cooler has no downside to it no matter how they want to spin the narrative. To present it as somehow being "optimized" to be functionally ideal while running 95°C is like something out of an NVIDIA playbook that is facetious even if their supporters choose to believe it. Getting a CPU, GPU or memory to run cooler never has any downside to it, and the cooler you can get it the longer it will last and the better it will overclock. There is no headroom available at 95°C. It is too close to the thermal junction maximum. One of the videos posted earlier actually demonstrates the problem. You can barely push the envelope and it goes into thermal shutdown mode. If you lower the temps by 20°C by correcting the engineering defect with a delid you are in much better shape in terms of overclocking potential than you were leaving it stock. In Roman's example the CPU automatically boosted higher all by itself and pulled fewer watts by dropping the temperatures to ~70-75°C with a delid. Edit: We see a similar example of this with DDR5. The manufacturer says is it "good" with 80°C temperatures. They ignore the fact that the memory starts throwing errors and the system becomes unstable at 45-50°C and you have to either lower the clock speed or lower the temperature to regain stability lost by higher temperatures.

This is an area where they would do well to follow Intel's lead and stop worrying about silly people whining like bratty little girls about compatibility with old parts and just do whatever is best to make the new stuff work as well as possible. If it is not reasonably compatible because it can't be, it is better to focus on achieving the best outcome and let the chips fall where they may on the idea of re-using old parts. Making the IHS too thick to fill the gap to achieve the same z-height at the expense of the CPU running ridiculously hot is kind of silly. Saying 95°C is the new normal as an excuse is equally silly... and a disingenuous distraction from the fact that it is hotter than it should be, and that impairs performance.