Mr. Fox

I do not tune anything with a goal of maintaining low power and low voltage. I always go for max performance and the power and voltage turns out to be whatever it is when I am done. I have not made any effort to tune the memory at 1.100V, so I have no idea. It certainly would not surprise me if a good bin of Hynix A-die could run 6000 with 1.100V with default timings. When you start cinching down the timings is when the low voltage begins to peter-out. You lose stability and can't boot anymore. That is why SO-DIMM modules usually have atrociously sloppy loose memory timings like pathetic 3200 CL20 DDR4 modules. That allows them to run stock low voltage and accommodate CPUs with garbage memory controllers. Well, it is good that the underlying code is there if you can get it unlocked and exposed. I have no idea what any of it means as it relates to the hex values. I am assuming the columns with zeroes reflect the absence of code on the modules if that is a dump from a live system. I am running 1.475V for 8000 CL35-48-48-42 2T.

That is good. They need that fan. The stock memory speed doesn't matter if you have access to memory controls and voltage. None of my DDR5 modules have XMP profiles. They are cheap generic server modules with no branding. The memory I have running at 6400 has a default of 4800 and the memory I have running at 8000 has a default clock of 5600. And, both have sloppy loose default timings. Manual memory overclocking is always better than XMP, but not if you cannot access all of the timings and voltage is locked.

It is unfortunate that there is such a profound chasm between what is allowed with desktops and laptops by the control freaks locking everything down. I am sure they are capable of achieving a massive degree greater performance than what is permitted through lockdowns. My Strix Z690-E 4-DIMM mobo really sucks at memory overclocking and it still runs 6400 CL30-38-38-28 2T and very tight secondary timings without breaking a sweat using Hynix M-die. And, my 24/7 memory overclock on the Dark is 8000 CL35-48-48-42 2T with tight secondary timings (Hynix A-die). The challenge would be keeping the memory from overheating in a laptop because overclocked DDR5 runs stupid hot in a desktop as well with the chintzy trash heatsinks heating blankets they put on them that look pretty and do a marvelous job of trapping heat. (Actually runs cooler naked sticks with air circulation than with the heating blankets.)

It should unless the memory maker has replaced the generic foundry markings with their own proprietary identification. Some of them do that, which kind of sucks. I have seen desktop and laptop RAM where they have obscured IC identification this way. But, you can use CPU-Z, AIDA64 or the shareware/demo version of Thaiphoon Burner to identify the IC if the markings have been stupified into something meaningless. Thaiphoon Burner won't let you manipulate DDR5, but it does tell you more about the IC and SPD firmware than any other software does. Of course, nothing is the same when it comes to laptops and SO-DIMMs and crippled firmware, so take my desktop memory experience with a grain of salt. It may be worthless. On desktop DDR5, Samsung B-die is on par with Hynix M-die until you try to push the clocks past 6200-6400 (varies by silicon lottery) and up to that point I found the Samsung B-die to be very stable and performant. I would assume (which is also worthless) that the Samsung B-die would be a good option for Ryzen since the maximum overclocking limitations are similar for the CPU and RAM. I was able to run Samsung B-die up to 6200-6400 in the Unify-X with really tight timings and low latency.

Hmm, well that did not help even a little bit. No change. I decided to install the first release beta driver for 6900 XT and so far that seems like the best yet. I also chose the minimal installation option with the most basic configuration options. Will see how it does over the next few days. I don't need any of the worthless gamerpunk stream-sissy bloatware or performance tuning capabilities, especially on my work OS. Is the IC Hynix A-die, M-die, something else, or you don't know until you get them? Even if you can't increase the frequency, if the laptop CPU has a half-decent memory controller you should be able to drop the primary timings to something substantially lower, like CL32-34-34-28 2T. Samsung should be able to do similar timings, as it is only limited in frequency headroom. Micron hasn't been great for DDR5 performance tuning.

It should as long as the latency doesn't get extra sloppy. Are these SO-DIMM modules in a laptop without an opportunity for memory tuning? If so, no idea how it might turn out, or if you can do anything to improve the outcome.

Don't cinch down the screws with the contact frame. It doesn't need to be tight. I've had to straighten bent CPU socket pins before. Both times were on the same motherboard and I could not identify a reason for the pins getting bent. This was on 10th Gen and no CPU frame. I think what happens is sometimes the pins can randomly not compress at exactly the same rate and may collide with one another while battening down the hatches. The speed at which the ILM arm is lowered (if too fast) might contribute to it happening. The bent pins can be hard to spot and only visible with light shining at an angle to help you see a difference in reflection. In the example I mentioned above, both times the bent pin was on the outer edge row, which might also have been caused by setting the CPU into the socket starting from an edge and "hinging" it down into place rather that setting it into the socket flat. I never had the problem before and never after. Being on the outer edge made the bent pin even harder to identify, There was less to contrast the light reflection against since it was not surrounded by other pins, but being on the outer edge row made it super easy to correct, pushing it back into alignment using a sewing needle. I did not have to worry about disturbing pins surrounding it on four sides.

No, as a matter of fact, I disable all "features" with AMD and NVIDIA because they offer me nothing of value and generally hurt benchmark scores. Everything I can find to disable is disabled. I just downgraded to their June 2022 driver (was on October 2022) and if that does not fix the problem, I may nuke everything with DDU, reinstall and omit CCC/Adrenaline all together and just install using the INF for barebones. The OS where this matters is the one I use for work and I do not need to adjust anything the control panel makes function. All I need is display output with no nonsense. I do not remember which GPU, I think the 2080 Ti, had lots of weird bugs with cheap DP cables and I had to buy expensive 8K high bandwidth CableMatters DP cables for my triple display setup. That fixed the problems I was having back then. I cannot use a 24-inch or shorter cable because it is not long enough to reach the monitor the furthest away from the GPU. I need at least 4 or 5 feet long.

Does it only affect I/O involving the drive(s) that are disappearing? I was about to say that did not sound right to me either. Especially not if the SSD is SATA or PCH connected. It is very unlikely the CPU is damaged or there would be other more significant signs. I would be leaning more toward the issue being something wrong with the operating system (or buggy storage driver) and the most likely culprit being actual drive itself having issues. Well, dog-gonnit, that sure didn't last long. The problem returned with a fury today. Trying to work and responding to emails between meetings became quite difficult because of disappearing text and artifacts/gibberish while typing. I confirmed the driver did not get a cancer update from the Redmond Reprobates. Same driver that helped 4 days ago is working as crummy as the latest. Very goofy. Several days of totally normal behavior, then suddenly the problem returns. One thing that I did observe immediately prior to the problem returning, and probably related, is that all of my desktop icons became black squares. I fixed that by deleting the files in the icon cache and restarting Windows Explorer. All of the icons immediately returned to normal, but the old issue returned. Weird. I've going to try doing a clean re-install of the AMD display driver to see if that re-fixes this font rendering glitch. Or, maybe I will try an even older driver based on this person's comments. I am on 22.20.27.07.

I know exactly what you are talking about. As a very giving, and forgiving person, I am sometimes taken advantage of by shady characters. Some of them belong in prison or something worse than that. That is not due to lack of wisdom as much as some very dishonest people are very good at being deceptive and dishonest. They are professional thieves and some are career criminals. Unless you have a crystal ball or simply treat all people as if they are dishonest and unworthy of your trust or respect, guilty until proven innocent, it is going to happen once in a while in spite of diligence and prudence.

I do not think he changed user name. Mobius was a different person that was upset and "after him" as in "pursuing" in a bad way.

Hi Brother! Long time no see. Doing better than I deserve to be. Nice to see you. Yes, what Brother @Papusansaid. It is what I use. Install only driver, Phys-X and HDMI audio if you use that (otherwise not) and use the tweak options to remove telemetry, GeFarts Excoriance, etc. If you skip the trash payload it will spare you about 1,200 registry entries, close to a dozen performance leeching Services and spyware that serves no useful purpose to the end user (only data piracy for the Green Goblin).

Brother @johnksssSo, yeah... ran a few loops of Time Spy with the 4090 stock and the reason it feels like a blast furnace with the air cooling is because it is. One could easily turn off their heater in the winter... 54°C (125°F) air getting pumped off the GPU within about 3 minutes, LOL. The arrow is where I have the temperature probe positioned. Definitely will improve with a water block for that reason alone. I don't know how hot it would get if I let it finish all 20 loops in the stress test. Public Service Announcement And, here is a tidbit for anyone that thinks that using alcohol to dry out the water inside of a water block is a smart thing to do. DON'T DO IT. I learned a valuable lesson the hard way. I flushed the 3060 Ti block with distilled water, poured some alcohol into it, maybe a tablespoon or two, swished it around, then blew it out with my PC cleaning blower. It removed all of the water quickly and effortlessly, and it looked like brand new. A few days later the acrylic is cracked severely and ruined. Now I know why my 2080 Ti Quantum Vector block is ruined. I did the same thing to it when I was getting ready to sell it. It looked like brand new. A few days later I got ready to take photos of it for selling and it was cracked like crazy. I hope my idiotic bright idea that is actually abject ignorance saves some heartache for one of you guys. Seems smart enough, but it is the exact opposite. In both cases the alcohol was in the block for less than 1 minute. That's all it took to destroy them both, and it wasn't obvious because I did not see it happen. It was not instant. I am so glad I did not use alcohol on the window of the DG-86 case. That would have been the end of it.

What about a reservoir and pump combo designed to fit into a 5.25 inch drive bay? These use a standard D5 pump. https://modmymods.com/reservoirs/bay-reservoirs.html This one even includes the pump. Price is actually really good... $87 USD. https://modmymods.com/alphacool-eisfach-single-laing-d5-incl-alphacool-eispumpe-vpp755-dual-5-25-bay-station-reservoir-13304.html

The EKWB Velocity2 direct die kit left a lot to be desired in terms of results. It was built like a Swiss watch, with impeccable precision and quality. But, the cooling results just were not there. The 6900 XT I got from Brother @Rage Setincluded a Bykski block to install and it does an impressive job, and looks nice. It doesn't have the same level of spit and polish to it, but it does what you want it to really well. Kind of like the Supercool direct die block. You can overlook a certain degree of crudeness based on results. I think the lack of reviews on the EKWB blocks might have something to do with results. Just speculation. If they were doing impressive things with thermals, I think we would see more. I say that knowing we live in a culture that tends to value form over function. That is sad, but it is the reality of the situation. The EKWB blocks are oozing form in spades. The function is not known due to the lack of reviews.

I should have it by this weekend. It is coming from Alabama (not China as is usually the case with Bykski parts). Based on what I saw with the Liquid X compared to the Strix and Suprim X (non-liquid) I have now, I think it will make a minor improvement with the chiller running. When I had the AC unit blowing cold air through the radiator on the Liquid X my boost clocks did go up, along with the benchmark scores. I anticipate similar with the chilled water. Hopefully a little better with the chiller. Probably no huge change in performance without the chiller, but absolutely better for me as the PC operator being in the same room as the computer in terms of the insane amount of heat generated by the air cooler. It is absolutely crazy how hot that air cooler gets and the heat bellowing off of it. Same was true with the Strix. In both cases the GPU stays respectably cool, but it is like heater running in my office when it is already too hot. The EVGA DG-86 case has a temperature probe for the digital read-out on the front panel. It is very accurate and always matches the digital thermometer I have sitting inside of the case. I put that probe into the outer end of the cooler. Maybe later I will run a Time Spy stress test and take a photo showing the temperature on the coldest end of the cooler. Should be interesting just for giggles.

I got a similar email to do a product review and I did a very negative review on the 4090 Strix because of its sucky memory overclocking. For some reason (LOL) they did not publish the review. 😉

They send it to you by email when you register an ASUS product to your account.

Thank you. Building it easy with flexible tubing and has the added advantage of making changes and redesigning on the fly easy. The hard part for me is keeping a system looking nice because of the absolutely brutal dust here. It's not just normal dust, but sand, so cleaning requires that you take extra care to avoid scratching things up really bad. And, no matter how much effort goes into it, eliminating scratches completely is borderline impossible. Just blowing the dust off with air can cause fine scratches on things like acrylic. Cooling systems are also less efficient here due to the lack of humidity. While it is helpful to have low humidity with a chiller, or something exotic like phase change or cascade cooling, in terms of avoiding condensation; in other circumstances (normal air or liquid cooling) the dry air makes radiators far less efficient. When the air has some moisture it adds an evaporative quality that really helps radiators (liquid and air) work better... kind of a swamp cooler effect, I suppose. During monsoon season, my system runs notably cooler than it does during normal Phoenix weather.

Sounds like a fantastic weekend with great results. That glue is probably going to be fine. The most important thing is that it is gel, not liquid. Ordered a Bykski block for the Suprim. Also redid my loop with new fittings and have it ready to go. I added a drain tube for the manifold to make it easier to drain. That black heat sink is just sitting on the back plate to help wick off some of the heat. And, it helped. That Suprim back plate gets crazy hot. Way hotter than the one on the Liquid X did, but that makes sense considering it is air cooled and going to get hotter. At least it is pulling the heat off the VRAM.

It still does not answer the question of why ROG Community--a product line that is supposed to be for enthusiasts--is catering to smartphone dumbphone jockeys and tablet craplet dummies. This says several things about ASUS that are not very good. Their focus has been lost. They should not give a rat's butt what the kiddos want. They should cater to the people and products the ROG brand markets to and tell the people that want to troll their forum on a phone to grow up and stop expecting everyone to pander to their childish whims. But wait... LOL... there are ROG smartphone dumbphones now. This is no shortage of filth for stupid people. Where is my barf bucket?

I'm glad that you like it. I find it much uglier, slower, more restrictive and resistant to modification than prior versions of Windows. Windows 10 after v1809 and all versions of Winduhz 11 are probably the least desirable versions of Windows since Millennium Edition. It's probably good that you skipped Windows 7 or you'd probably strongly dislike Windoze 10 and Winduhz 11 and find it very difficult to let it go like I do. They haven't really done anything good or right since Windows 7 in my opinion. All they have done is make it slower, uglier, added more data harvesting malware and increased the payload of worthless filth. Other than tweaking a few unnecessary services, Windows 7 is ready to rock out of the box. W8.X and everything that came after have progressively required greater amounts tweaking and tuning and hacks in order to correct aesthetic atrocities to make them palatable for me. I hate having to waste my time fixing their mistakes, bad judgment and poor taste in GUI design.

Thank you. It comes as a relief to me, too. (That is kind of a "Captian Obvious" comment on my part, LOL.) I hope it doesn't just mysteriously die like the other one did. Two good ones in a row for me, but this one is a little better. I think the liquid cooler tend to do something to hold back the Liquid X because all of the online reviews show the non-Liquid generally edging out the Liquid version by a tiny margin. They are exactly the same GPU with a different cooler. Maybe MSI is putting the better ones on air and the lesser binned ones on water to help level the playing field. I haven't heard any coil whine yet. The heat this air cooled monster puts out is just ungodly. It definitely needs a block to help contain that. https://hwbot.org/submission/5235384_ Pre-benchmark starting temperatures... I was getting about 5K more than this on Fire Strike with the Galax HOF vBIOS, so I will be going there again soon. Running these now as a point of reference to confirm improvement later after cross-flashing.

It is totally unacceptable. The GPU manufacturers are using whatever parts come to them and letting the people that buy them figure out if they got sold a piece of garbage. If that's how they are going to roll, the price should be half what they currently sell for. Consumer dishonestly has always been a problem. Even during good times without a lot of trash tech floating around. The behavior we see from the hardware OEMs and lack of QC and concern for customer experience fuels dshonesty where it would never otherwise enter a person's mind. The difference now is the OEMs deserve to get screwed back... real hard. Now it is more like consumers returning the favor, tit for tat. I'd almost condone it rather than discourage it at this point, because they're screwing us deliberately. And, you know what they say about payback... bow-wow, doggy-girl.

Stock MSI vBIOS. Will move to the Galax HOF soon. Cold air is keeping core max to 40°C and core max to 48°C with max fans and boosting to 3150 (+225) and +1650 memory offset. https://www.3dmark.com/3dm11/15686404 This excites me almost as much as if it were my own experience. Super happy for you. The delid is long overdue. I anticipate your thermal issues will be largely a thing of the past now. Heating up the IHS is probably what allowed you to push off the IHS without damaging/denting the side of it with the push block of the delid tool. I don't heat the IHS. I will try that the next time. How did you warm it up? I have not delidded the 13900KF in my work computer because I don't ever push the CPU overclock hard enough to thermal throttle on that machine. 5.7GHz on P-cores is as high as I have ever pushed it and that is my daily driver clock on the work machine. Maybe I will try heating it up when I delid that CPU and see if it avoids denting the side of the IHS.