Mr. Fox

It sucked when they ruined Battlefield and Call of Duty single-player campaign with this rancid feces. I refuse to purchase games where the developers have gone down the dictatorship path. I wouldn't install and play them even if they were free. I am happy to say that I don't care about Fartnite so this butthead move won't affect me directly, but it still sucks because some people do care about it. I really do hope that all of these game developers start losing money and struggle to maintain solvency because of their stupid decisions. They do not deserve to be successful when they force anti-cheat cancer on people that only play single-player campaign. #36,537·a moment ago (Edited) I will not purchase or play (even if free of charge) any game that requires me to enable TPM or Secure Poot in the BIOS. I have been buying and playing Battlefield and Call of Duty since the first release in each franchise, but I am done with them now. I only play single-player campaign. Yet, they insist on dictating this unacceptable filth even in circumstances where cheating is a non-issue, so to hell with them and their cancer game crap. They will not receive another cent from me if this is how they roll now. And, I am not the only former player they have lost. If they were paying attention they would recognize that this dictatorship mentality is one of the reasons that Micro$lop is losing OS market share at an unprecedented rate. PC gamers and enthusiasts are not Kool-Aid drinking zombie sheeple like crApple users are.

I think they know that they have already lost the support and confidence of all of the smart people of the world. That leaves them with people that are not smart, which buy crap like Arm devices and disposable trashbooks.

https://www.overclock.net/posts/29559222/ So I did more testing and this is where I am at now. My max boost is no longer hitting 5950 and maxing out at 5819. I will see if I can figure that out. Maybe I can deepen CO a bit more on one or two of my best cores and get that back. But, no more TM5 or Karhu errors. Passes fine now with PBO same as it did with manual OC. So, it was CO causing the errors and not memory. I also ran two iterations of CoreCycler without error. (Man, talk about boring... takes forever, LOL.) What is interesting is how the individual core SP ratings in the BIOS coincide with the stable CO negative offset. So far I have found cores with the same SP rating are stable with the same negative CO offset. I am sure there is some fine tuning to be done on top of that, but it might be a good starting point or point of reference in troubleshooting. With the current CO settings (5819 MHz vs 5950 MHz max clock) CPU-Z single thread shows a loss of about 8-10 points and multi-thread about 200 points loss. If I use "Balanced" power profile and enable core parking all core intensive benchmarks end up in the toilet. This is a Windoze garbage OS performance-destroying issue (Micro$lop calls it a "feature" LOL) and not hardware.

After all is said and done, with Ryzen the benefit (or lack thereof) doesn't scale with the effort involved. Lots of time and fussing to get it right with a teeny-tiny (or no) gain in return. Nothing to write home about going from 8000 to 8400. Exactly the opposite with Intel memory overclocking.

OK. Enough of the silly gamerboy normie crap. Back to speed trumps everything.

If you keep the GPU for a few years it will be worth doing. When you to move to a newer platform the GPU will be ready for it.

Dropping tRFC to 388 shaved a ns off of memory latency, but otherwise unremarkable by other measurements. There is very little difference between 6400 1:1 and 8000/8200 2:1 on Ryzen due to architectural limitations. Adding 3D cache makes it even less relevant. Y-cruncher is around ~1.0 second slower at 6400.

How plentiful are waterblocks for the Red Devil? When I still had the AORUS 9070 XT the only one was an Alphacool block that was never available for purchase. It was always pre-order and/or out of stock. That played a significant part in my decision to get rid of it. I was sick of living with a hot air-cooled GPU. It was miserable. Radeon GPUs take "hot" to the next level of cray cray when it comes to air cooling.

You better snag that 3090 Ti while you can. Hopefully it is not a scam. If it is not that is an insanely good deal. If you are not going to buy it, I will.

Which 9070 XT do you have, Brother @Rage Set. (I think you mentioned it before but I do not remember.) I got bored and decided to mess with slower memory speeds to see how fast I could make it go. This is stable so I will see how low I can get tCL before it unravels. 6400 1:1 is snappy feeling even though it is a bit slower on read/write/copy.

PCWorld generally is not a good resource for useful technical information. I am not interested in the biased thoughts and personal opinions of their editors. I only care about what I think and seeing how many others share my opinions. It used to be a good source, but not for a long time. They do not allow readers to comment on their reviews and articles, which delegitimizes them in my opinion. I think any media outlet that does not allow readers to comment to support or rebut has an agenda of some sort and they want to control the narrative. I posted my thoughts on the TweakTown article yesterday.

I thought this was a good (and accurate) assessment of the situation. We are getting screwed just because the industry is run by opportunistic trash scumbags and they are doing it just because they can. What goes around comes around and payback is going to be a real b**** for all of them before its over. Once the illogical AI slop build-out has run its course they are not going to have a product to sell because there is will be not enough demand for it at the inflated price and they'll have to roll back prices to mid-2025 levels or go out of business. AI is not bad. AI development companies are bad. Their motives and methods are wicked and they will reap what they are sowing.

There is never a shortage of insanely overpriced trash made for the dumbest people in the world. A year or two ago I would not have believed there would be a human being dumb enough to pay that for a mouse, but now I know they are real and there are enough of them to be dangerous to all of the people or ordinary or greater intelligence. https://www.overclock.net/posts/29556653/

The Thermaltake Core W200 was an awesome case. I have always wanted one. In the rare circumstance I could even find one they were so ridiculously priced that I could not justify the cost. There have only been a couple that I am aware of. Even monster cases like the Corsair 1000D Obsidian only support mini-ITX for the second mobo and SFX for the secondary PSU. The Phanteks Enthoo 719 also only supports mini-ITX for the second mobo, but does support two ATX PSUs. I had a Thermaltake Level 20 XT E-ATX Extreme LCS Capable Split Level Modular Gaming Cube Case CA-1L1-00F1WN-00 that I purchased in 2019. It supported dual ATX PSUs and you could add a second ATX mobo tray. Those are also nowhere to be found. It was not crazy tall, but it was so huge from front to back that the feet would not fit on either of my desks. I had to put it sideways on the desk. It was my favorite case of all time except for the fact that it needed a dedicated desk just for the case. It was so gigantic that I just could not accommodate it in my tiny home office. If I had a bigger office I would never have gotten rid of it. It was amazing. You might just want to go with a server rack setup with two 4U chassis units that each support ATX/E-ATX boards, ATX PSUs and full sized GPUs. Leave the lid off (for GPU power cable clearance if necessary) and leave an empty space between them in the rack. Maybe a couple of units like this: https://www.newegg.com/rosewill-rsv-l4500u-black/p/N82E16811147328?Item=N82E16811147328 and mount them in a rack. https://www.newegg.com/p/2BA-01UW-00010 A more expensive option (maybe not worth it compared to the Rosewill) would be two of these: https://www.newegg.com/p/2AM-006F-001D2 mounted in a rack. Silverstone also makes an RM51 5U chassis, but you would be looking at $1,000-$1,200 (at the least) for a pair of them and a rack. https://www.provantage.com/silverstone-rm51~7SLST033.htm?source=googleps

The question is, "Are there really going to be massive profits?" or are they spending a lot of money on a gamble that may or may not eventually pay off? I am not convinced they are going to make money and may end up drawing back a bloody stump before it's over. I certainly hope so. It would be a real shame if their cash-grabbing nonsense turns out well for them and for their shareholders. I'd much prefer to see financial tragedy visited upon them. Tit for tat.

The least they could do is pretend they love us and say nice things while they are choking us and pulling our hair.

Even the Quadro P5000 MXM in my Dell Precision turdbook has 16GB of vRAM. It is inexcusable that 8GB GPU options exist today. That should only be on the cheapest and lowest tier garbage-class products that are just one step up from integrated graphics. Nothing new is built to last the test of time or deliver maximum performance. Everything is designed to fail, and deliver structured performance limitations, to ensure a never-ending flow of cash. It is remarkably similar to the pharmaceutical industry and practice of medicine. They do not make money on healthy people or cures. Keeping people on meds forever, and follow up doctor visits to renew prescriptions, and creating new medical problems from prolonged consumption of chemicals that require more meds and more medical care is the gift that keeps on giving. They don't want anyone to be healthy because it is bad for business. PC and electronics hardware companies do not want to manufacture and sell durable consumer products that perform well. They want to sell garbage with a limited lifespan so they can collect more money on upgrades and product refreshes to replace the last generation of trash they sold consumers. Live to pay, pay to live. Just like the old bumper sticker... Ass, gas or grass... nobody rides for free.

I was going to post this video. I'm glad I saved money going with the Zotac 5090.

They are all part of the same circle jerk and worshipping at the same Satanic throne of AI filth.

Yeah, that should be fine. They also constantly change under load to a small degree. That is also normal. If you start seeing 2.0 to 3.0 amps variance under load where one pin is always that much lower than the rest of the pins then you should start investigating what the deal is because something is wrong. You got a lot of stuff done. I look forward to hearing how all the other things worked out once you have more time for testing. If you had a poor mount that would certainly diminish your memory overclock. You do not need to have the high performance IHS super tight. Only tight enough that it fully presses the CPU into the socket pins and the screws are not loose.