GTXJackBauerYou could play around with the fan curve.  I think your temps are fine.  It's when maybe they might start baking and rising up after hours of gameplay you might not notice.  Again, these GPUs can handle up to 88c-89c and the GPU will protect itself via thermal shutdown.  If the Vram gets really warm/hot, that could start artifacts as well and cause crashes, etc.  I've seen many show high temps in those areas but in order to have more temps sensors across the PCB, you would need to have a , and as those are the only ones that come with the iCX2 sensors that give you GPU, VRAM and VRM temps.

AHowes
That fan adapter would only serve well for a blower style card. Your card is sucking air from below and blowing out against that fan blowing in and threw the sides of the card. Prob better then nothing I guess since it's the only thing helping pull air in as the weak 200mm fan is blocked from the drive bays.

Newer style cases dont have drive bays blocking the front fans anymore as they made room for 360/480 radiators.

You might be able to move the front drive bay if you can move the drives up top.

I have an older half x case that was better for air flow as it had a side fan blowing right at the card.. but now they dont do that anymore with the glass panel trend. Same 200mm fans in the front and top. Fine for an older pc.. used it for a media pc.

Also used it for a while on my sons pc with a 140mm fan blowing cool air directly at the card.. kept a 1080 classified nice and cool.. and before that a pair of 980s.

So if that case has a steel door or plexy glass door maybe cutting a hole for a 120/140mm fan or 2 will no doubt help a ton.

11-05-2019, 21:53:08 TheDivision2.exe benchmark completed, 132952 frames rendered in 2244.234 s
                     Average framerate  :   59.2 FPS
                     Minimum framerate  :   29.6 FPS
                     Maximum framerate  :   60.5 FPS
                     1% low framerate   :   29.8 FPS
                     0.1% low framerate :   28.6 FPS

jasoncodispoti

jeffyt
Small update, played some Division 2 last night for about 2 1/2 hours with the last 30 min running MSI Afterburner benchmark. Temps were around 73-75 on the GPU around the 2 hour mark. The clock was around 1890-1950 and I also dropped one setting, Volumetric Fog to low, which resulted in ~20% GPU usage drop and I didn't see the 131-149% power spikes anymore. I think it was mostly around 90-110%, some 120% spikes.
 

11-05-2019, 21:53:08 TheDivision2.exe benchmark completed, 132952 frames rendered in 2244.234 s
                   Average framerate  :   59.2 FPS
                   Minimum framerate  :   29.6 FPS
                   Maximum framerate  :   60.5 FPS
                   1% low framerate   :   29.8 FPS
                   0.1% low framerate :   28.6 FPS

Did notice a tiny amount of stutter since I use vsync and it looks like I averaged 59 fps, so I'll drop nvidia control panel settings to Performance instead of Quality for texture filtering and hope that'll get the 60 average as I've never liked to run any OC on the GPU at all.

What resolution are you running at? Why do you use vsync? 

jeffyt

jasoncodispotiWhat resolution are you running at? Why do you use vsync?

AHowes
He's using 4k at 60hz thats why the v-sync.

Yes 4k at 60hz, I can't stand the screen tearing. I've heard about scanline with RTSS where you can hide the tear off screen but they say it's only good for 60-70% GPU usage. Anything higher and it still tears normally I think.

I've had some more luck with RTSS FPS limit which seems to have stabilized the frame timing/pace but still averaging the 59 fps unfortunately. 425 drivers have a smoother feeling, but it's been a short test since DDU and installing 425, so it may be placebo for now. Plus I'm a little worried about security so I'll update to the newest 430 drivers soon after some more testing.

The description for Event ID 13 from source nvlddmkm cannot be found. Either the component that raises this event is not installed on your local computer or the installation is corrupted. You can install or repair the component on the local computer.

If the event originated on another computer, the display information had to be saved with the event.

The following information was included with the event:

\Device\Video3
Graphics SM Warp Exception on (GPC 5, TPC 1, SM 0): TEX FORMAT Errors

The message resource is present but the message was not found in the message table

jasoncodispotiOh that explains the low frame rate...

GTXJackBauerYou will want to stay @ 59/60 FPS since that's all your screen can handle at refreshing.  Anything beyond that isn't visible on your screen and only uses up your GPU for nothing.  If you want higher FPS, you'll need a screen with a higher refresh rate than 60hz.  You want to match the Hz with FPS.

jasoncodispoti
What version of Windows are you using? 
 
Edit: That dump file that you provided indicates that the blue screen occurred due to a DPC_Watchdog_Violation which typically indicates some type of hard drive issue, most of the time firmware or driver issue... Investigating stand by. What SSD drive do you have? Or HDD? Or both?

AHowes
Well if that's the only change and its stable now. Its gotta be another bad card.

Keep testing.. and contact evga again about it and hopefully without you even suggesting it, they send you a new card this time.. if not, ask for one.

Have they even tested the cards your sending back to report one what they found?

bcavnaughThis is with your Motherboard Set to Auto on the CPU and Memory or are you Overclocking your CPU and Memory on your Motherboard?
 
"clarity: 1st card: newegg, retail packaging. 2nd card: evga, retail packing 3rd/4th/5th: evga, RMA re-certified/refurbished non-retail very basic packaging."
So you are now on your 5th Card?

jeffyt

jasoncodispoti
What version of Windows are you using? 
 
Edit: That dump file that you provided indicates that the blue screen occurred due to a DPC_Watchdog_Violation which typically indicates some type of hard drive issue, most of the time firmware or driver issue... Investigating stand by. What SSD drive do you have? Or HDD? Or both?

Honestly, I appreciate it but I feel like if the only change I made was -200/-300 on clocks of GPU that fixed it... I mean.. how could it possibly even be anything else?
 
Anyways, I'm running Windows 10 and I have a Samsung SSD 850 EVO 500GB as main windows/game drives. With two storage HDD, 8tb and 3tb.  WDC WD30EZRZ-00WN9B0 and WDC WD80EFZX-68UW8N0

1. You will need to go into the BIOS and reset all settings back to factory defaults and them leave them unchanged. 
2. Update the BIOS to the most recent version, 7A62v17 https://www.msi.com/Motherboard/support/Z270-GAMING-M3#down-bios. 
3. Re-install Windows 10, complete clean install of the OS using the Windows 10 media installation tool found here, https://www.microsoft.com/en-us/software-download/windows10. (may want to create the media prior to doing steps 1 and 2)
4. Once the install is complete you need to install all of the applicable drivers for the motherboard which can be located here, https://www.msi.com/Motherboard/support/Z270-GAMING-M3#down-driver&Win10%2064.
5. Reboot PC.
6. Install Nvidia Drivers. 
7. Reboot PC.
8. Check for and install all Windows 10 updates. You will need to keep repeating this step until you no longer find additional Windows updates to install. 
9. Download and install the Samsung Magician software, open the software and confirm that your SSD health is GOOD and that its running the latest firmware, software can be located here https://www.samsung.com/semiconductor/minisite/ssd/product/consumer/magician/.
10. Download and install OCCT found here, http://www.ocbase.com/index.php/download. After you have installed OCCT you are going to want to run the OCCT test for 1 hour using a Large Data Set. This will help to confirm that the CPU, motherboard, and RAM are stable and not having issues. 
11. It would be ideal if you could test the ram to 98% coverage using Karhusoftware's RAM Test application, but this is a paid application and you can skip if you feel that you most. The application can be found here, https://www.karhusoftware.com/ramtest/.
12. Now we need to test the GPU using something like 3DMark's FireStrike Extreme on a 1 hour loop would be ideal, but once again this is a paid application. If you dont want to purchase 3DMark than go ahead and use FurMark and run the stress test for 1 hour. You can locate and download Furmark from here, https://www.techspot.com/downloads/5462-furmark.html.