Noticed some frame rate stutters and screen tearing.
Fire up HWMonitor. Observe 100% GPU usage, clock speed at 683MHz (boost clock can get up to like 1354 MHz, so it’s way low), and GPU fan at 0 RPM.
Pull out graphics card and blow out with compressed air. Fan bearings seem fine. Heard no fan noises leading up to this point.
Plug graphics card back in to see what happens. Fan spins briefly on startup, turns off, and never turns back on again. Didn’t run graphics card stress test, but was maxing out GPU usage again with a client running.
Use DDU and do a fresh install of the graphics card drivers. No change.
Follow the instructions here, in order to manually turn on GPU fan.
Set fan speed to 33% and it works!!! GPU clock speed bounces up to 1215MHz, and GPU utilization drops to normal.
Manually setting fan speed gets the job done, but was wondering why my graphics card stopped automatically controlling fan speed, and if it can be fixed. Looked through a few google search results, but I’m getting tired and brane knot werk sew gud. Thus, I thought I would ask to see if anyone had any insight. Plus, this thread might help others with fan/overheating issues.
P.S. For anyone who might read this who still can’t get their fan to spin and are desperate, my next step was going to be to remove the plastic shroud around the heatsink, and direct a small box fan on the card. Might need to turn down the client’s graphics settings, but hopefully it will be enough to allow you to keep playing until you get a new fan or graphics card.
Also, if your fan spins, and you blew out the card with compressed air, but the card is still hot and throttling, you probably need to replace the thermal paste.
Could just be a defective GPU… When I bought my new PC back in Aug, The GPU would start having issues and turned out it was defective, and they sent me a new one…
How old is this Card ?
What Vendor ?
Is there a newer BIOS Version for your Card ?
Do you have a test Card ?
Install the Vendor Tool Software, often you can make own Profile’s for the CPU Temp / Fan Speed.
You use win10 ?
GTX 960 is an older Grafikcard, about’s 5-6 Years old.
For one Cient its okay, for more then two, the VRAM if the Card run low.
What Platform to you use ?
I expericed with an Intel i5 - 2600 perfomance problems with 2/3+ Clients.
Do you have the same problems with an fresh Installed Windows ?
Use GPU-Z and CoreTemp for Temperatures of you CPU and Grafikcard.
IDK, 5 or so years old. But it wasn’t used much for 2 or 3 years.
EVGA
Win10
Ryzen 5 3600 (stock speed) 32GB RAM
Windows installation is about 3 months old, and is malware and bloatware free.
zero problems with a GTX 1060
No issues with VRAM. It’s a 4GB card, and Eve doesn’t use that much (I once actually tested GPU memory usage and found that Eve uses between 84 and 300MB depending on client settings at 1080p when sitting on station undock. I would expect large fleet fights to use more, but I’d be surprised if I ever have any problems with VRAM, especially considering how I usually stay on potato mode).
lol, and I’m sure people will give me ■■■■ for not mentioning it earlier, but I didn’t feel like getting into an argument about whether or not you can run 2 different graphics cards, or whether or not that’s the source of my problem. But I suppose I would be remiss if I didn’t mention it. I run a 960 and a 1060. Anyway, I’ve seen people mention that running 2 different graphics cards is bad juju, but I’ve been running these two cards without issue for about 6 months now (first on an intel platform, and now on an AMD. I also ran the 1060 with an AMD card on the intel platform for about 2 years before that). Moreover, there are many other people who also dispute that running 2 different cards is bad juju. It may cause problems in some cases, but it is not an inherently bad solution. Moreover, my problem started yesterday.
Speaking of which, I had not made any hardware changes, setting changes, driver updates, software updates, or anything prior to the problem starting (I manually update everything except for my anti-virus because I don’t want a bunch of updater processes constantly running in the background).
GPU temp has stayed between 26 and 41 degrees celcius now that I have the fan manually turned on.
Yeah, I installed the vendor tools in order to manually set the fan speed. Does look like I can set my own fan speed curve like you mentioned through, so I might try giving that a shot.
I think that should answer all your questions, now for an update.
After I manually set the fan speed, I ended up experiencing more frame rate drops, and noticed max utilization and a throttled clock speed again. This was in spite of my GPU temp being about 30 degrees celcius at the time. I read somewhere that some people can have problems with their clock speed “sticking”, and to change their GPU power management mode from “adaptive” to “prefer maximum performance.” So, I’m going to give that a shot this evening.
I’m also going to re-test my PSU. It tested well within specs when I upgraded my system about 3 months back, but it’s also 8 years old, and has seen extensive use.
Speaking of which, I actually checked 2 different PSU calculators before my upgrade. I don’t remember the exact numbers, but my PSU had over a hundred watts over the recommended wattage. So, that shouldn’t be an issue.
And if you’re wondering why I run 2 different graphics cards, it’s because I got both of my secondary cards for free when friends/family upgraded. So, I decided to use them to drive different displays, and then break up the workload as much as possible. For example, I multibox Eve, so I’ll split the clients between the 2 cards. Or, I’ll run a game on one card/display (with settings cranked up as much as I can), and have video playback/web browser/discord/maps/whatever split across two other displays on the other card. Now, I will probably upgrade to a beefier card within the next year, but this has given me extra gpu muscle without spending any more money.
… hmm… what NV Driver Version du you have ?
So, you have the Problem since yasterday, i would test the Card alone in an test system.
Because, i think you cant use differnt NV Driver for Differnet Grafikcard Generations.
Every dirver Versionen is normaly adjust on the latest Hardware.
So maybe there is this Problem. But this is only a theory.
To end sticking around, best way is to test the 960 in an testsystem alone, to watch how it is doing, if you have the Hardware at home.
PSU Problems are like, hangups and freezes if the 12V is getting to weak.
For Normal 650W should be okay for this system.
More i can’t say. If you need two card, watch out in “willhaben.at” for a 1060.
With the new 6800 Cards and the 3080 Card, this Chrismas should be easy to get used Hardware.
I don’t have any more used Hardware … i equiped some people in my corp with my old stuff XD
I use a B450 + Ryzen 5 2600 + 2060 Super with 32GB RAM and i have only 550W.
I run 5 EVE Clients + Web + Netflix on 3 Display same Time.
An the System is running stable.
Well, it appears to be working now. Ran for a couple of hours with power management set to “maximum performance” without any issues. Decided to switch back to “adaptive” to see what would happen, and ran for 3 hours without any issues. So, hell if I know. Maybe it was a hiccup due to having two different cards, maybe the card is failing, maybe it was something else. IDK. Regardless, if I encounter the problem again, I’ll probably just switch back to maximum performance and leave it there.
Anyway, like I said, I plan on upgrading within the next year, so I just need to hold out a little bit longer. Fore one, I’d like to see what’s up with the RTX 3060 TI and RTX 3060 before making a decision, and two, I’m not paying inflated prices to scalpers. So, I’ll wait till availability improves.