Mr. Fox

That absolutely sucks. Unforgivable. The only end users they are treatin right is RTX 6000 owners. Bastards. I wonder how long it will continue to be available from the Micro$lop Store Whores? This also sucks...

At this point it certainly seems that the quest is to identify the least undesirable option because none of them are good. The best plan might actually be to chose nothing and reject all of the options. What I would love to see is an aggressive wiper malware that specifically targets and destroys all AI source code in all platforms across the globe, and that it would spread like wildfire, uncontrollable and utterly destroy all of it and financially obliterate the companies backing it. Google Gemini, Chat GPT, Grok, Claude, LLaMA, CoreWeave, Baidu, Qwen, Tencent, ByteDance, AWS, Accenture... aIl of it FUBAR'd and unrecoverable, along with the money wasted on it... totally destroyed, permanently corrupted, unusable and nothing salvageable, and no resources available to rebuild with.

That sentiment has not changed. It was too much then and still too much now. The current prices are untenable and any gamer paying current prices for a new 5090 GPU is bereft of common sense. They either have so much money they have lost the ability to think rationally about their purchases, or they should not be trusted to make any important decisions about anything that requires average intelligence. Radeon remains rough around the edges and hard for me to get overly excited about as an overclocking enthusiast, but 9070XT is undeniably the best bang for buck gamer proposition that exists. That excludes any models sporting the petite arson connector. Those should be viewed as undesirable trash. It defies logic that they were able to sell any of them. The most compelling reason for buying a Radeon GPU is the avoidance of that trashy abortion power connector.

Have you tried using Lossless Scaling for 3DMark benchmarks to see if it works?

Not as hard as justifying the purchase of a 5090. The cost of two brand new is still less than the cheapest 5090. The 5090 offers less value per dollar than any other PC component in the history of personal computers. Current pricing is reserved for the mentally insane.

I love it when big tech companies suffereing from AGS (AI Glorification Syndrome) fail and lose gobs of money. Good job, Micro$lop. I hope it hurts more and causes more long-term damage than they are willing to admit.

@tps3443 since you just did a clean install, you might as well do it again. The latest release of Rufus nukes a ton of Micro$lop's bundled feces prior to installation. https://rufus.ie/en/ Changelog Version 4.14 (2026.04.30) Windows User Experience improvements: Add a Quality of Life option, to disable Teams, Outlook, Copilot and other Microsoft forced nuisances Add a Silent installation option, that automatically, and WITHOUT PROMPT, installs Windows on the first detected disk Add an option to copy SkuSiPolicy.p7b to the ESP on installation (please refer to KB5042562 for more info) Add tooltips for all the dialog options Add limited support for El-Torito UEFI image extraction (Mostly for Dell BIOS update ISOs) Improve error report when the user tries to use an image that resides on the target drive Improve the UEFI:NTFS partition label to make the install media more explicit during Windows Setup disk partitioning Improve support for Bazzite and other Fedora derivatives that don't follow EFI conventions Improve detection and exclusion of the new Bitdefender hidden VHDs Improve reporting of GRUB and Isolinux MBRs Fix potential errors during creation of Windows To Go media, due to the use of new versions of bcdboot Fix errors with local accounts that start or end with whitespaces

Rufus update just rammed an ice pick into the gonads of the Winduhz 11 Cancer OS. But, it's still a piece of crap and smart people choose something better. https://rufus.ie/en/ Changelog Version 4.14 (2026.04.30) Windows User Experience improvements: Add a Quality of Life option, to disable Teams, Outlook, Copilot and other Microsoft forced nuisances Add a Silent installation option, that automatically, and WITHOUT PROMPT, installs Windows on the first detected disk Add an option to copy SkuSiPolicy.p7b to the ESP on installation (please refer to KB5042562 for more info) Add tooltips for all the dialog options Add limited support for El-Torito UEFI image extraction (Mostly for Dell BIOS update ISOs) Improve error report when the user tries to use an image that resides on the target drive Improve the UEFI:NTFS partition label to make the install media more explicit during Windows Setup disk partitioning Improve support for Bazzite and other Fedora derivatives that don't follow EFI conventions Improve detection and exclusion of the new Bitdefender hidden VHDs Improve reporting of GRUB and Isolinux MBRs Fix potential errors during creation of Windows To Go media, due to the use of new versions of bcdboot Fix errors with local accounts that start or end with whitespaces

I am glad it was something "expendable" that is not affecting your core system. Enjoy that nice Zotac 5080 Solid OC GPU now. That was a good one to buy. Are you keeping both or sending back the new one?

I bet that was the problem. I hope so. If that fixes it I suspect you are going to find some other things that are mysteriously working better than they used to. You might even find games and YouTube videos seem to run smoother. Failing USB devices have caused me headaches more than once. Sometimes subtle glitches and sometimes outright malfunction of the entire system. I think they must sometimes develop a short internally. I also had a USB webcam that developed an issue that caused weird problems when it was connected.

Ah ha! Crossing my fingers and hoping that does it. Very possible. That would be a very good problem to have if you are going to have one. I bet the USB hub inside of the monitor has a problem. I used to have an Acer Predator 1440p monitor that developed a similar problem after a couple of years of working fine. Even with nothing connected to the ports, just the cable from the PC to the monitor, it was frequently freezing and making the Windows hardware disconnect and reconnect sounds. It was randomly changing between working to "code 43" in Device Manager and causing stuttering in games and just web browsing or playing YouTube videos. Once I disconnected it everything worked fine. What clued me in was the Windows sounds alerts for hardware disconnect and reconnect and when I opened Device Manager and saw it happening it was instant confirmation.

What if you put your son's GPU in your system and run the same test again? Is your system PCIe Gen 3 or Gen 4? Did you try to manually set the max for that in the BIOS already? It is interesting that it seems to require the passage of time for it to occur. Like something has to reach a certain point. I am wondering if maybe one of your memory sticks is bad. Maybe try one at a time and see if one is messed up. Since you ruled out the NVMe, if I were going to take a wild guess at this point, it would be one of the sticks is bad and the system works until data reaches that one stick and then it's over. That's were I would start at this point. Might not hurt to check the temperature of the memory modules at the point of lockup. Probably won't see anything but checking never hurts. That is a good suggestion. It might show something relevant.

Hopefully that is it. If you only had a single drive attached to your system that will make it easier to test for sure since you only have the one potential culprit. I had a 4 TB TeamGroup NVMe that failed back when I still owned the Z790 Apex. It was causing lockups and before I could figure out what the cause was it got worse and wouldn't even let the motherboard power on. It was like it was unplugged from the wall. Pressing the power button did nothing. No signs of life. As soon as I removed it everything returned to normal and I had it replaced under warranty.

I wouldn't rule out that the problem could be caused by one of the drives. If you have multiple SSDs or HDDs any one of them can cause lockups like what you are experiencing. If the one you replaced doesn't fix the problem don't stop there. You need to test all of them. Particularly the one that the games are installed on. Failing drives, even USB storage, can cause all sorts of weird problems including hard lockups. I have even seen defective internal and USB drives that prevented the PC form powering on and failed booting even at the BIOS level and hard lockups are a very common symptom.

Have you tested with only one video card in the bottom slot instead of the top slot where you would normally use a GPU? Sorry to hear this is happening but also glad that the second hand 5080 you bought wasn't defective. It's probably safe to assume that you have tested both video cards or at least the one you bought on Facebook in your son's computer and there are no issues. It is a long shot but you may want to try the GOP firmware update released by Nvidia recently. I applied that update to both my 5080 and my 5090. It is possible the old UEFI/GOP firmware on your motherboard is not playing nice and that might help. The landing page on the Nvidia website where I downloaded it gives the impression it is for the 5060 only. That is misleading because it applies to all 50 series graphics cards. When I flashed it the nvflash scanned the hardware ID to confirm that the update was applicable before proceeding. As it turns out it was applicable to both of my video cards.

For gaming very little benefit since there is no SLI hardware support and almost zero SLI support in terms of game development. You could use one for render and one for compute, and you could use Lossless Scaling for gaming. That would be a very expensive second GPU for that though. What GPU does your son have? If he has anything less than a 4090 it would be a nice upgrade for him. If you wanted to experiment with Lossless Scaling, you could use his GPU for that as a second GPU in your system. All of the above aside, having two matching 5080 Solid GPUs installed in your system would look pretty wicked. If you end up returning the new one you will have an "original owner" invoice with your name on it to register for warranty with Zotac. You could buy this one 5090 for roughly the cost of 4.5 5080s if you wanted to do something extra-stupid and frivolous with your money. https://www.overclock.net/threads/msi-rtx-5090-lightning-z-65-of-1300-custom-loop-compatible.1819519/?post_id=29587541#post-29587541

Yeah, the Zotac hate is really ignorant. They make great GPUs. Their warranty sucks, but the same can be said of other brands. A$$zeus and Gigabutt look for excuses to deny warranty for totally illegitimate reasons. A warranty that isn't honored because a company is dishonest is the same as having no warranty. Tony from Northwest Repair has mentioned in several videos that the Zotac GPUs are very well made. I think he has even said they are the best built. They have fuses to protect them and some of the more popular and more expensive brands do not.

Yay! That's great news. As much as I love my 5090 and I am glad that I have it, at current prices I would not even consider buying one again. The 5080 is a little weaker than a 4090 but out of the three options it is the only GPU that makes any sense financially. If you can even find a good used 4090 they are selling for prices that are just about as stupid as 5090 prices. I really like my 5080. I didn't feel dirty or like an idiot after buying it based on the price.

Sounds like switching the cable might do the trick if you are passing a SN stress test. Much further than you had gotten before from the way it sounds. Were you able to look at the PSU end of the 12VHPWR cable? If that adapter is working it could be the cable or the PSU, or maybe a bad connection on the PSU end of the cable (or melted at the PSU end).

This is exactly why I urged testing it in your son's computer. It is certainly possible that there might be something wrong with the 12VHPWR cable or the PSU. I would also look closely at the PCIe slot to make sure there is not some fuzz or other debris inside of the slot that is blocking contact with the contacts on the GPU. If your motherboard is PCIe 4.0 I would also try manually setting the PCIe speed for the GPU slot to whatever its maximum spec is rather than leaving it set to Auto. Are you going to try to return the new one, or give your son a nice upgrade?

@Papusan 2.0