Yeah My company only just rolling Win 11 out, we spent last year testing/re-writing internal Apps. Some were still 32 bit Everyone still bracing themselves…
I have a hunch that this may be the GPU + mobo problem.
I was using the igpu for the entire day and the system boots flawlessly.
Then I installed the proart rtx 4070 super oc and after reading about the 12v cable adaptor problems etc I paid in attention to plug everything right and firm.
The only way (at least tha I am aware of) to replicate the fail boot (with the RTX GPU installed) is to wait a very certain amount of time between shutting down and booting up. Its about a minute or two. Then between the white q-led and the green boot q-led will be a significant time-gap and with a bit of luck the pc will boot or the boot q-led will never lit and pc will freeze. If I wait longer like around 20-30 minutes the pc will boot up just fine, as well as when I turn it on just a few seconds after it turns off.
As of now I try to fail-boot with the iGPU but with no success. It seems to boot up just fine no matter what.
What are your system specs exactly?
What you are experiencing during cold boot couldn´t that just be the memory training process?
What you could try is to go into the bios.
And ¨enable¨ the ¨Memory Context Restore feature.¨
It should be somewhere in the ai tweaker menu under dram timing settings.
ryzen 9 9900x
kingston fury 2x 32gb ddr5 at 4800 speed not overclocked
rtx 4070 super oc proart
proart x670-3 mobo
nzxt 850w gold psu
arctic liquid freezer III 420mm AiO
fractal north xl case (it may gonna be relevant since Ive checked a vast majority of things already)
answering your question
Im stuck after all q-leds, the dram that goes first, then cpu, then vga and after vga fades the computer outputs no signal to displays, it freezes and in an act of despair (keeping in mind that this may be some wicked ram training issue) I left it in that state for 7 hours, just went out to work. Nothing changed, it was constantly bricked. Pressing power button twice (rebooting) instantly fixed that.
I was suspecting the DP cable that would have the 20th pin wired, but I just confirmed its a vesa certified cable with 20 pin not wired. I checked it using voltmeter.
Right now I’m testing the RTX on 1 display plugged by HDMI only. It seems to be stable so far, I could not trigger the failed boot.
I ordered another proart rtx 4070 to see if it may be the issue. I should have it in 2 days…
Ok. I dunno of this was really looked at, but I thought I would throw in a idea since you seem ok with parts swapping.
I have a ASUS Rog Strix X670E-E and a 7900x3d with a RX 6950XT.
I had boot issues, memory stability, random reboots, and so much more. Only some Bios worked, overclocking of ram or gpu was impossible. I would shut it down and get 00 on the bios read out and all the fans lights etc would stay on…yeah funky.
After chasing my tail for a long time I found it was a faulty or rather unclean voltage from the PSU causing things to go haywire… new PSU… no problems since. I forget what voltage was the issue (maybe 5), and because it was unclean messed with clocks and memory voltages and cpu voltages. Nothing got fried it just keep hitting the protectuin values and would reboot, shut down or hang.
Just and anecdotal experience but thought it might help because I dont think I saw PSU mentioned.
Now I can use any bios I want, overclock ram, use power curves etc and sytem is 100% stable at any power state.
The most frustrating is that I cannot cause this fail boot on demand. Even stripping my rig to nothing and teying to bood adding pieces will never give me an answer because it takes hours in a several config/state to finally happen…
Like I wrote above that I can cause this issue just by waiting 1-2 minutes between the boots - I find it no longer valid since for the last several hours Ive been doing any kind of routine that would have caused the boot to fail previously, but now it boots just fine. I have literally no idea, I just wasted the whole day, I am tired and frustrated.
Have a good day everyone!
Thanks for your hint, I have replaced this psu a year ago and had absolutely no problems with it on my previous platform, where I was using i9 10900k, gtx 1080ti ftw3 64fb ram and gigabyte visionD mobo. The previous setup was quite comparably energy consuming, plus I had my i9 overclocked to 5.1ghz.
It was rock stable, never ever failed so I assume that nothing could have happened to this psu… Or this may be a sorry coincidence.
I thought the same thing. I had a Corsair AX1200i. No issues at all with the last build till I moved over to the x670 platform. I dunno if it was more sensitive to voltage or what but my issues were totally random as well. I could not do anything to say “doing this makes it happen”. Could work for hours being beatup by tests, and then fail opening a browser window lol.
Just a thought.
Damn you gave me a serious food for thought…
Either way i hope you find a solution. No fun having your rig down.
I just found a thread about the DP issues that I may be having. Some people and perhaps manufacturers claim, that having an old DP cable that is a few generations older than the new GPU’s output signal, may cause among others a hang on a boot. My DP Cables are quite old (at least 7 years old) and I did not choose them conciously to be some ultra high-end cables, I just wanted them to be VESA certified at the time. They may be dp rev 1.0. What you guys think?
I have seen this on older systems. Some kind of “power issues” might be involved. (not necessarily PSU, maybe mobo power thingies)
When You “rapidly” power cycle, energy stored in capacitors does not have enough time to dissipate. Sometimes devices require more power for just a moment when turning on.
Yeah My x670E doesn’t restart properly. half the mobo still alive, when it does makes a popping sound like the capacitors are throwing a wobbly. Few posts said remove sleep options and full shutdown pull cable a few times and try and restart then, seems to trick it into full restart. Really been a pain in the butt tbh
I did not mean that the only way to boot whatsoever. I could make some time gap between and still boot, it’s just a matter of another attempt.
UPDATE:
It’s been over 36 hours since the last failed boot, I booted my system at least 200 times.
What has been done:
- I’ve undone the cooler screws by a quarter turn
- I am using hdmi and DP cables combo (I assumed that at least 1 DP cable may be faulty).
- I use only 2 m.2 drives, I put away the third one that was chipset operated
- I made sure to wire the GPU flawlessly with no loose on the plug (the problematic 12V adapter from NVIDIA)
- I attached the optional CPU 8 pin power cable
- I ensured that my DP cable doesnt have 20th pin wired
The bios rev 2007 seems to be stable so far. I see noticeably higher cpu temperatures with this bios, but as long as the PC is booting I cannot complain (not that I purchased one of the most expensive motherboards on the market…, it should at least cook me a dinner.)
Is there a difference between rebooting using windows menu and rebooting via physical reset key?
My solution which seemed to get it stable was to fully remove Armoury Crate with Revo uninstaller, dog piece of software leaves shit everywhere.
Good to hear that it works
I did not mean to imply that rapid-powercycle is the only way, but over the years I have seen systems that required it at first sometimes, and then it got worse.
Also, I’ve always considered 8-pin cpu power cable as mandatory.
Shutdown using Windows… then pull the plug hit the start button to clear the capacitors then plug in and reboot.
do you mean on boards, that have 2 x 8-pin connector on top left? Even the user manuals say that one of them is (always?) optional, just for sake of heavy overclocking