Also I realized they aren't identical, the other one is a titanium 750 and this one is a platinum 750.
That's efficiency, so it matters only for your power bill. How much power the PSU is able to deliver to the components depends on the ampere rating, most important on 12V. Still worth trying, no PSUs are identical, some get slightly better parts, some get worse parts, even if within the tolerances. And than of course some simply wear out faster then others for no obvious reason. Even if you buy high quality stuff.
From the looks of it you are dumping on the new GW tasks which put a much bigger load on both the cpu and gpu. No issues with the much easier GR and Meerkat tasks. And of course the Separation tasks at MW are a trifle for any gpu.
How were you able to tell that? I'd love to know for future troubleshooting.
Thanks for the advice everyone. I was just hoping to exhaust my free options first (driver problem, Windows settings etc) since a Seasonic or EVGA 1000w+ titanium PSU are easily $300+.
I turned kombo boost back on and found it actually reduced my CPU power usage from ~100w to 88w while boosting clock speed so kudos to MSI for figuring out how to pull that off. I also rearranged the cables on my power supply since it's fully modular thinking maybe there was a faulty plug. Neither made a difference, computer restarted after just a few minutes of running Einstein. Sounds like it's not worth pursuing measuring the computer's power draw. I'll start thinking about my next move.
By looking at the output files of all you errored tasks.
But I was looking at the wrong computer it seems. The majority of your GW errors are on the 5800X host. I see that your OP was about the 5800X3D host.
That computer seems to be throwing errors on just about all types of work. Lots of illegal instructions errors. That points to memory errors likely and may be because of the bigger L3 cache on that chip. Possibly when you updated the OS Windows bungled handling the software package for properly configuring the cpu.
Still could be inadequate power supply. I agree that swapping power supplies between the two hosts would be the fastest way to troubleshoot.
But I don't know if that is wise since the host with the supposedly "good" power supply is showing it is not very stable either. Way too many errors. Lots of storage system errors where files can't be found or files are corrupted and don't match their CRC checksums. Check that the storage system is configured for the proper bus speed and width.
Either way it seems you are pushing both systems way too far with overclocking or whatever. I would revert both systems to stock out of the box configurations for both memory and cpu clock speed. No MSI Kombuster in use please. See if you can get one day of accurate calculations with no errors.
By looking at the output files of all you errored tasks.
Is that something I'm able to do as a run of the mill user or does it require special access?
As far memory errors I have a spare 2x 8GB kit I can throw in. I know you suspect the CPU cache but couldn't hurt to try.
The 5800x computer has been having a problem lately where the "boinc screensaver loading" icon floats around but it never actually loads. I just found out 2 days ago that if I suspended milkyway it loads and runs Einstein just fine. Maybe because of that storage system error you found. Stange because the hardrive I store boinc data on works fine otherwise. It ran Einstein pretty much all day and all night. As soon as I resumed milkyway, infinite floating icon again. For now though I want to focus on getting the 5800x3d pc figured out and I'll continue to run only Einstein on the 5800x pc.
The 5800x3d pc ran milkyway all night last night. Unfortunately I'm working all weekend so won't be able to start swapping parts until Monday.
Either way it seems you are pushing both systems way too far with overclocking or whatever. I would revert both systems to stock out of the box configurations for both memory and cpu clock speed. No MSI Kombuster in use please. See if you can get one day of accurate calculations with no errors.
Yes, getting it stable at stock speed/timings/whatever is definitely first thing to do, the error rates are way too high.
Did some more troubleshooting today. I had the computer running milkyway all last night and today with no crashes. I checked the errors for this computer and there are none since I suspended einstein, except for the tasks that expired. I ran the built-in Windows 10 memory diagnostic tool and it found no errors. Next step is to run Memtest86.
I also swapped power supplies. The "good" 5800x computer now has the platinum 750 and the "bad" 5800x3d computer has the titanium 750. The 5800x computer ran einstein with the platinum power supply with no crashes. The 5800x3d computer still crashed with the titanium power supply. This time however, I got a blue screen instead of abruptly restarting. The first time I wasn't expecting the blue screen so I ran einstein again and managed to snap a picture. When I went to open the minidump file, it was filled with text symbols instead of coherent english. I used wordpad and notepad to open it. Can anyone tell me what I'm missing here?
Along with running memtest86, I'm going to use display driver uninstaller to wipe the AMD graphics drivers off the computer, then install the drivers from December of '21. Those are the drivers I'm using on the 5800x computer and have been working fine. Can you guys think of anything else I should try?
Did some more troubleshooting today. I had the computer running milkyway all last night and today with no crashes. I checked the errors for this computer and there are none since I suspended einstein, except for the tasks that expired. I ran the built-in Windows 10 memory diagnostic tool and it found no errors. Next step is to run Memtest86.
I also swapped power supplies. The "good" 5800x computer now has the platinum 750 and the "bad" 5800x3d computer has the titanium 750. The 5800x computer ran einstein with the platinum power supply with no crashes. The 5800x3d computer still crashed with the titanium power supply. This time however, I got a blue screen instead of abruptly restarting. The first time I wasn't expecting the blue screen so I ran einstein again and managed to snap a picture. When I went to open the minidump file, it was filled with text symbols instead of coherent english. I used wordpad and notepad to open it. Can anyone tell me what I'm missing here?
Along with running memtest86, I'm going to use display driver uninstaller to wipe the AMD graphics drivers off the computer, then install the drivers from December of '21. Those are the drivers I'm using on the 5800x computer and have been working fine. Can you guys think of anything else I should try?
I would do a check of the hard drive as long as you are doing testing stuff, it should happen pretty quickly and hopefully won't find anything but you never know. Oh and I'm assuming you blew everything out when you were swapping psu's but if not it's not a bad thing to do.
The 5800x3d computer still crashed with the titanium power supply. This time however, I got a blue screen instead of abruptly restarting.
Did you finally downclock that thing to stock? THAT is the first thing to do since it's also crashing with the 2nd PSU, everything else comes later if it's still not running stable. I'm pretty sure it will also crash with those clocks if you run Prime95 on CPU and Furmark on GPU, you are pushing it simply too high.
The 5800x3d computer still crashed with the titanium power supply. This time however, I got a blue screen instead of abruptly restarting.
Did you finally downclock that thing to stock? THAT is the first thing to do since it's also crashing with the 2nd PSU, everything else comes later if it's still not running stable. I'm pretty sure it will also crash with those clocks if you run Prime95 on CPU and Furmark on GPU, you are pushing it simply too high.
Yes I disabled kombo strike and disabled smart access memory in bios. I posted a screenshot of the blue screen error, is anyone able to interpret that? I feel like that could certainly send me in the right direction.
I have also ran Memtest86+ to test the RAM some more. Windows built in memory tester didn't find any errors but this one found over 60,000! I ran it on the "good" 5800x PC which has 4x8GB sticks and it passed. I'm going to throw in my spare kit of 2x8GB and try einstein.
The 5800x3d computer still crashed with the titanium power supply. This time however, I got a blue screen instead of abruptly restarting.
Did you finally downclock that thing to stock? THAT is the first thing to do since it's also crashing with the 2nd PSU, everything else comes later if it's still not running stable. I'm pretty sure it will also crash with those clocks if you run Prime95 on CPU and Furmark on GPU, you are pushing it simply too high.
Yes I disabled kombo strike and disabled smart access memory in bios. I posted a screenshot of the blue screen error, is anyone able to interpret that? I feel like that could certainly send me in the right direction.
I have also ran Memtest86+ to test the RAM some more. Windows built in memory tester didn't find any errors but this one found over 60,000! I ran it on the "good" 5800x PC which has 4x8GB sticks and it passed. I'm going to throw in my spare kit of 2x8GB and try einstein.
Traditionally those error messages are just memory dumps with little discernable value to most of us.
Personally I would run the Memtest86+ on the new ram before you run Einstein, just to rule that out if the tasks fail.
Adam wrote:Also I realized
)
That's efficiency, so it matters only for your power bill. How much power the PSU is able to deliver to the components depends on the ampere rating, most important on 12V. Still worth trying, no PSUs are identical, some get slightly better parts, some get worse parts, even if within the tolerances. And than of course some simply wear out faster then others for no obvious reason. Even if you buy high quality stuff.
.
Adam wrote: Keith Myers
)
By looking at the output files of all you errored tasks.
But I was looking at the wrong computer it seems. The majority of your GW errors are on the 5800X host. I see that your OP was about the 5800X3D host.
That computer seems to be throwing errors on just about all types of work. Lots of illegal instructions errors. That points to memory errors likely and may be because of the bigger L3 cache on that chip. Possibly when you updated the OS Windows bungled handling the software package for properly configuring the cpu.
Still could be inadequate power supply. I agree that swapping power supplies between the two hosts would be the fastest way to troubleshoot.
But I don't know if that is wise since the host with the supposedly "good" power supply is showing it is not very stable either. Way too many errors. Lots of storage system errors where files can't be found or files are corrupted and don't match their CRC checksums. Check that the storage system is configured for the proper bus speed and width.
Either way it seems you are pushing both systems way too far with overclocking or whatever. I would revert both systems to stock out of the box configurations for both memory and cpu clock speed. No MSI Kombuster in use please. See if you can get one day of accurate calculations with no errors.
Keith Myers wrote: By
)
Is that something I'm able to do as a run of the mill user or does it require special access?
As far memory errors I have a spare 2x 8GB kit I can throw in. I know you suspect the CPU cache but couldn't hurt to try.
The 5800x computer has been having a problem lately where the "boinc screensaver loading" icon floats around but it never actually loads. I just found out 2 days ago that if I suspended milkyway it loads and runs Einstein just fine. Maybe because of that storage system error you found. Stange because the hardrive I store boinc data on works fine otherwise. It ran Einstein pretty much all day and all night. As soon as I resumed milkyway, infinite floating icon again. For now though I want to focus on getting the 5800x3d pc figured out and I'll continue to run only Einstein on the 5800x pc.
The 5800x3d pc ran milkyway all night last night. Unfortunately I'm working all weekend so won't be able to start swapping parts until Monday.
You and anyone else can look
)
You and anyone else can look at your returned results as long as your computers aren't hidden.
Just open up the errored tasks page on the website and click on the task result link.
Then read through the result looking for the error that was logged for the task.
You might also want to run a memory scanner and see if it passes without throwing errors.
Keith Myers wrote:Either way
)
Yes, getting it stable at stock speed/timings/whatever is definitely first thing to do, the error rates are way too high.
.
Did some more troubleshooting
)
Did some more troubleshooting today. I had the computer running milkyway all last night and today with no crashes. I checked the errors for this computer and there are none since I suspended einstein, except for the tasks that expired. I ran the built-in Windows 10 memory diagnostic tool and it found no errors. Next step is to run Memtest86.
I also swapped power supplies. The "good" 5800x computer now has the platinum 750 and the "bad" 5800x3d computer has the titanium 750. The 5800x computer ran einstein with the platinum power supply with no crashes. The 5800x3d computer still crashed with the titanium power supply. This time however, I got a blue screen instead of abruptly restarting. The first time I wasn't expecting the blue screen so I ran einstein again and managed to snap a picture. When I went to open the minidump file, it was filled with text symbols instead of coherent english. I used wordpad and notepad to open it. Can anyone tell me what I'm missing here?
https://imgur.com/Y5p6GuB
Along with running memtest86, I'm going to use display driver uninstaller to wipe the AMD graphics drivers off the computer, then install the drivers from December of '21. Those are the drivers I'm using on the 5800x computer and have been working fine. Can you guys think of anything else I should try?
Adam wrote:Did some more
)
I would do a check of the hard drive as long as you are doing testing stuff, it should happen pretty quickly and hopefully won't find anything but you never know. Oh and I'm assuming you blew everything out when you were swapping psu's but if not it's not a bad thing to do.
Adam wrote:The 5800x3d
)
Did you finally downclock that thing to stock? THAT is the first thing to do since it's also crashing with the 2nd PSU, everything else comes later if it's still not running stable. I'm pretty sure it will also crash with those clocks if you run Prime95 on CPU and Furmark on GPU, you are pushing it simply too high.
.
Link wrote:Adam wrote:The
)
Yes I disabled kombo strike and disabled smart access memory in bios. I posted a screenshot of the blue screen error, is anyone able to interpret that? I feel like that could certainly send me in the right direction.
I have also ran Memtest86+ to test the RAM some more. Windows built in memory tester didn't find any errors but this one found over 60,000! I ran it on the "good" 5800x PC which has 4x8GB sticks and it passed. I'm going to throw in my spare kit of 2x8GB and try einstein.
Adam wrote: Link
)
Traditionally those error messages are just memory dumps with little discernable value to most of us.
Personally I would run the Memtest86+ on the new ram before you run Einstein, just to rule that out if the tasks fail.