- Joined
- Mar 18, 2015
- Messages
- 2,960 (0.89/day)
- Location
- Long Island
1. If you look at other posters messages, under their avatar you will see "system specs", please put you full specs in there; also consider that some of us old folks, after many years of PC use, (and teenagers for other reasons) have RSI to one level or another and if the "main info" isn't out there in plain view, some will just move on. That's your an their perogotaive, but recognize it may cut down on number of responses.
2. Knowing what make / model of the card (as well as everything else) is helpful.
3. Always include max CPU / GPU temps and voltages under load
4. Addressing the thermal issues on a card is a common solution and one of the few "universal" causes regardless of model. For example, if it was an EVGA SC 970, I'd say ... is it one of the early ones where 1/3 of the heat sink "missed" the GPU. Or if it was a EVGA 1060 - 1080, then Id ask if you know is it was one of the early ones where they cheaped out and left out the thermal pads. No way to tell if replacing TIM or reinstalling cooler will help as there is no information indicating temps at 100% load.
5. I understand that taking a component in your computer and taking it apart may be daunting. And you certainly have the right to put whatever label on it you want; just recognoze that 99% of the world calls this "troubleshooting" or "maintenance". My son just replaced one of the "chips" on his car (224,000 miles) . He had the choice of having the dealer do it ... $900 for the part + $300 labor or do it himself. He found one online from a junkyard 2 sates away for $20 and will install it himself, after having a local dealership program it for him ($75). You have the choice of returning the card and having them perform the "mutilation" or do it yourself. This is no different than having a car with a bad spark plug or blown water pump gasket. The fact remains, these products are made on production lines that move very fast. Not every one gets built to ideal standard of care. Very often a card survives w/o needing such attention ... if it wasn't fastened down evenly, it can deteriorate. Not also that the card came out almost 4 years ago and no one at the company from the penthouse office to low level line worker cares whether the card lasts past it's 3rd year. While it's noy a 100% thing, having to replace your TIM on CPU / GPU is by no means an unusual thing after 3 years. Think of it like the engine coolant on your car.
6. A power cut, in and off itself, may not be an issue. The problem occurs if there was a power surge before it went out or when it came back on. When power goes out, if you don't have adequate protection (UPS or Surge Protector), always unplug your PC. A $20 power strip w/ "surge protection" on the label is not good enough ... If it's less then $75, always unplug ... if it's more than $75 ... should unplug anyway ... risk may be small (0.01%) but unplugged it's 0%. Everything in your house isn't on at the same time .... but when power goes out for any length of time, temps drop in house, temps rise in fridge and when they are called to go on .... they can't. So when power is restired, the big instantaneous suck of power results in a voltage drop which means amps rise. Protection of your PC downstream of any protective devices starts with your PSU ... we don't know what that is so can't examine that failure path intelligently.
7. Asides from hardware impacts, the loss of power can case data loss or corruption. in order of easiest to hardest, things you may have to examine include:
a) Uninstall and reinstall drivers ... BTDT
b) Do a clean install of new drivers ....
- Uninstall all things nVidia from control panel
- Clean registry of all things nVidia using a registry cleaner.
or
use TPUs nvida uninstaller / installer utility
https://www.techpowerup.com/forums/...an-installer-for-nvidia-drivers-alpha.249085/
or
Use DDU
https://www.guru3d.com/files-details/display-driver-uninstaller-download.html
then reinstall all items in proper order
8. While the logs may be useful, w/o knowing the specific card or the boost range permitted in MSI AB, troubleshooting is limited by a lack of information. Which specific 980 Ti are we talking about ? I saw that CPU temps question was asked ... did I miss the results ?
9. I note that the power usage seems to have dropped even further since you pushed the limit up. One of the problems here is a varying load during the test as it's hard to run 2 tests thru the exact same sequence and lining up / comparing comparable results. Before helping someone with troubleshooting, I usually ask they run something that is sure to keep CPU load and GPU load up as close as we can to 100%. Also noted that your 3770 is the hottest running Intel desktop CPUs we have seen in last decade. This is what I use to collect this data. If you prefer other tools, by all means. Other tools will serve, just avoid P95 or other sytnhetic type CPU tests
https://www.hwinfo.com/
http://dlcdnmkt.asus.com/rog/RealBench_v2.56.zip
https://geeks3d.com/furmark/
a) Launch HWiNFO (sensors only) suggest setting it up to span screen top to bottom on right side of screen
b) Launch RoG Real Bench and run on left side of screen so you can see HWiNFO results during the 8 minute test. Record temps and voltages in each core as well as start and finish times.
c) Run Furmark ... we are not looking here to check GPU OC stability, you are not OCing so the goal here is to see if the card is able to get high numbers on % load. Furmark will show you temps on screen. Stop test when temps flat line or if ya see anything scary. Record temps and power usage as well as start and finish times.
d) record cause of any failures or crashes
e) Open event viewer and record and warnings or errors during times that tests were being run.
10. Again, specific information on all products in play is the best hint at what may be the problem. Is this a reference card, or AIB ?...who made it and what "model line". Same for PSU
So far ...
MoBo: Maximus IV Gene-Z/Gen3
CPU: Intel 3770k
Ram: 24 GB how many sticks ? identical ?
GPU: Gigabyte Model ____ 980 Ti
PSU:
Storage:
OS:
As was said, some components gain a history of known problems due to experiences by other users which may hint at what your are expeiencing
11. Have ya tested your storage device integrity ?
12. Look for reporting errors in Windows Logs (Event Viewer)
13. Checked the rail voltages on your 3.3, 5.0 and 12.0 volt rails (HWiNFO) forgot to put this in item 9.
14. Run the usual OS health checks ..
Win7
sfc / scannow
https://www.thewindowsclub.com/fix-windows-update-using-dism
Win 10
https://answers.microsoft.com/en-us...ntegrity/75529fd4-fac7-4653-893a-dd8cd4b4db00
15. Of course sometimes hardware problems can only be resolved but switching them out and seeing if problem goes away. If none of the above points to anything specific and HW is ruled out, then a OS reinstall may be only solution.
If ya are post additional info, will try and take this further.
2. Knowing what make / model of the card (as well as everything else) is helpful.
3. Always include max CPU / GPU temps and voltages under load
4. Addressing the thermal issues on a card is a common solution and one of the few "universal" causes regardless of model. For example, if it was an EVGA SC 970, I'd say ... is it one of the early ones where 1/3 of the heat sink "missed" the GPU. Or if it was a EVGA 1060 - 1080, then Id ask if you know is it was one of the early ones where they cheaped out and left out the thermal pads. No way to tell if replacing TIM or reinstalling cooler will help as there is no information indicating temps at 100% load.
5. I understand that taking a component in your computer and taking it apart may be daunting. And you certainly have the right to put whatever label on it you want; just recognoze that 99% of the world calls this "troubleshooting" or "maintenance". My son just replaced one of the "chips" on his car (224,000 miles) . He had the choice of having the dealer do it ... $900 for the part + $300 labor or do it himself. He found one online from a junkyard 2 sates away for $20 and will install it himself, after having a local dealership program it for him ($75). You have the choice of returning the card and having them perform the "mutilation" or do it yourself. This is no different than having a car with a bad spark plug or blown water pump gasket. The fact remains, these products are made on production lines that move very fast. Not every one gets built to ideal standard of care. Very often a card survives w/o needing such attention ... if it wasn't fastened down evenly, it can deteriorate. Not also that the card came out almost 4 years ago and no one at the company from the penthouse office to low level line worker cares whether the card lasts past it's 3rd year. While it's noy a 100% thing, having to replace your TIM on CPU / GPU is by no means an unusual thing after 3 years. Think of it like the engine coolant on your car.
6. A power cut, in and off itself, may not be an issue. The problem occurs if there was a power surge before it went out or when it came back on. When power goes out, if you don't have adequate protection (UPS or Surge Protector), always unplug your PC. A $20 power strip w/ "surge protection" on the label is not good enough ... If it's less then $75, always unplug ... if it's more than $75 ... should unplug anyway ... risk may be small (0.01%) but unplugged it's 0%. Everything in your house isn't on at the same time .... but when power goes out for any length of time, temps drop in house, temps rise in fridge and when they are called to go on .... they can't. So when power is restired, the big instantaneous suck of power results in a voltage drop which means amps rise. Protection of your PC downstream of any protective devices starts with your PSU ... we don't know what that is so can't examine that failure path intelligently.
7. Asides from hardware impacts, the loss of power can case data loss or corruption. in order of easiest to hardest, things you may have to examine include:
a) Uninstall and reinstall drivers ... BTDT
b) Do a clean install of new drivers ....
- Uninstall all things nVidia from control panel
- Clean registry of all things nVidia using a registry cleaner.
or
use TPUs nvida uninstaller / installer utility
https://www.techpowerup.com/forums/...an-installer-for-nvidia-drivers-alpha.249085/
or
Use DDU
https://www.guru3d.com/files-details/display-driver-uninstaller-download.html
then reinstall all items in proper order
8. While the logs may be useful, w/o knowing the specific card or the boost range permitted in MSI AB, troubleshooting is limited by a lack of information. Which specific 980 Ti are we talking about ? I saw that CPU temps question was asked ... did I miss the results ?
9. I note that the power usage seems to have dropped even further since you pushed the limit up. One of the problems here is a varying load during the test as it's hard to run 2 tests thru the exact same sequence and lining up / comparing comparable results. Before helping someone with troubleshooting, I usually ask they run something that is sure to keep CPU load and GPU load up as close as we can to 100%. Also noted that your 3770 is the hottest running Intel desktop CPUs we have seen in last decade. This is what I use to collect this data. If you prefer other tools, by all means. Other tools will serve, just avoid P95 or other sytnhetic type CPU tests
https://www.hwinfo.com/
http://dlcdnmkt.asus.com/rog/RealBench_v2.56.zip
https://geeks3d.com/furmark/
a) Launch HWiNFO (sensors only) suggest setting it up to span screen top to bottom on right side of screen
b) Launch RoG Real Bench and run on left side of screen so you can see HWiNFO results during the 8 minute test. Record temps and voltages in each core as well as start and finish times.
c) Run Furmark ... we are not looking here to check GPU OC stability, you are not OCing so the goal here is to see if the card is able to get high numbers on % load. Furmark will show you temps on screen. Stop test when temps flat line or if ya see anything scary. Record temps and power usage as well as start and finish times.
d) record cause of any failures or crashes
e) Open event viewer and record and warnings or errors during times that tests were being run.
10. Again, specific information on all products in play is the best hint at what may be the problem. Is this a reference card, or AIB ?...who made it and what "model line". Same for PSU
So far ...
MoBo: Maximus IV Gene-Z/Gen3
CPU: Intel 3770k
Ram: 24 GB how many sticks ? identical ?
GPU: Gigabyte Model ____ 980 Ti
PSU:
Storage:
OS:
As was said, some components gain a history of known problems due to experiences by other users which may hint at what your are expeiencing
11. Have ya tested your storage device integrity ?
12. Look for reporting errors in Windows Logs (Event Viewer)
13. Checked the rail voltages on your 3.3, 5.0 and 12.0 volt rails (HWiNFO) forgot to put this in item 9.
14. Run the usual OS health checks ..
Win7
sfc / scannow
https://www.thewindowsclub.com/fix-windows-update-using-dism
Win 10
https://answers.microsoft.com/en-us...ntegrity/75529fd4-fac7-4653-893a-dd8cd4b4db00
15. Of course sometimes hardware problems can only be resolved but switching them out and seeing if problem goes away. If none of the above points to anything specific and HW is ruled out, then a OS reinstall may be only solution.
If ya are post additional info, will try and take this further.