Wish I could fully understand that command. I tried for hours.

sudo command is simple. Place it before you enter restricted commands, then enter your password. If you are in sudo user group it will work.

OK explanation of the options.
 
 nvidia-xconfig is the command you use to have the system configure your xorg.cfg file so the drivers AND Xwindows itself can find the cards and set them up for correct usage.
 -a   configure ALL cards in the system.
 --cool-bits=31     set up so you can do fan speed control, overclocking/underclocking, TDP control, and such.
 --force-generate    set up the configuration from scratch, and overwrite anything that is in the way if needed while moving any existing xorg.conf file to a "backup" version filename.
 --mode 1280x1024   tell Xwindows to start up in this screen mode, and default to it, if the card supports this mode on the port it has a monitor attached to.
 
 The number of dashes is CRITICAL - don't leave one out or add an extra one as that will mess up that configuration option.
 Don't leave out spaces on the options that have spaces, or add them to the options that don't have them - see above.
 
 
 nvidia-settings is a pain to figure out without examples, but it's how you control clocks and fan speeds and it gets rather arcane at times.
 As one major weirdness, you have to set the CPU or Memory clock in 2 slightly different spots for them to actually "take" - only change one spot, it doesn't "take", at least on 10xx series cards.
 I'll get into that in a later post, I need to do some FTP work to get an example over to this machine.

(NOTE: this config file only works if you have two GPU's that are identical. if you have more/less than two GPU's try creating your own with `sudo nvidia-xconfig -a --cool-bits=12 --allow-empty-initial-configuration`)

 
that is the easiest thing you can try to get coolbits enabled. if that doesnt work you'll need to try and figure out how to manually config your xorg.conf file. go here to try and learn a bit more about the process: https://foldingforum.org/...c.php?f=89&t=25075

Thanks - can't wait to get home and try this out
 
Might need to reinstall linux though...I tried using nvclock earlier and it seems to have made my display black with a grey bar at the bottom and I can't see how to recover it :/

GPU OC settings after reboot.
This works on Linux Mint.
I never was able to get it to work by hand editing xorg.conf. What did work was to execute on the command line which sets it all up for you. This worked on two of my builds so try this. On my third build it didn't work and I just copied the file from a build where it did work, and that worked. If this doesn't work for you I'll try and get you the file that works for me:
sudo nvidia-xconfig -a --cool-bits=31 --allow-empty-initial-configuration
Then edit xorg.conf. For me that was sudo vi /etc/X11/xorg.conf and prepend "#" to each line containing AllowEmptyInitialConfiguration to comment it out. If I didn't then I got duplicate icons in my panel. I have no idea why.
Reboot.
Then to overclock run:
/usr/bin/nvidia-settings
To restore your settings after a reboot create an executable file that you call from startup applications containing the text below which will set the gpu clock offset and set the gpu to prefer maximum performance. My example sets the offset to 80. Don't set the offset too high in the file for your actual display gpu until you know for sure what you want or you may end up with a system where the display won't work. FYI the accepted wisdom is that OC of memory doesn't help PPD but in my experience it does. YMMV:
nvidia-settings -a [gpu:0]/GpuPowerMizerMode=1 # 0=adaptive, 1=max,2 or none=auto
nvidia-settings -a [gpu:0]/GPUGraphicsClockOffset[3]=80
nvidia-settings -a [gpu:0]/GPUMemoryTransferRateOffset[3]=300
#to control GPU power execute the below example for GPU 0 and 110 watts. To get the valid wattages enter something like 1000 and it will respond with the valid values for your GPU
sudo -n nvidia-smi -i 0 --persistence-mode=1
sudo -n nvidia-smi -i 0 --power-limit=110
nvidia-smi # to just display current values
#I execute the below in my script that sets my overclock settings as nvidia-settings needs to be started after the OC is set.
sleep 10
/usr/bin/nvidia-settings &
 

On my system there is no xorg.conf by default and it gets created when I execute the command.  I've done it a dozen times and it works perfectly every time.  And speaking of fan control, I don't do it because the built in fan control works fine for me.  I don't want to go too low and overheat my cards or too high and wear out my fans prematurely, that being the only mechanical part and subject to mechanical wear. I believe the card designers know more about it than I do.  Plus gpu temperature varies with the project and I want my fans to adjust for that.

Unfortunately, recent card models almost ALWAYS set the fans way too low in their "default profile" to keep the things reasonably cool.
The whole EVGA "thermal pads as a fix" issue was actually more about that, as I HAVE some of the card models that were involved in that and never had an issue with them because I don't run the fans at the insanely WIMP stock fan profile.
I'll also note that EVGA changed the bios on them in later production to INCREASE the fan speed.
 
Card manufacturers want to balance NOISE vs HEAT - and tend to come down on the side of "more quiet" in recent years as a selling point, assuming that the large majority of their cards are run in gaming which doesn't put consistant high loads on the card causing consistant high average power draw.
For compute usage, THIS IS AN ISSUE as it tends to have the cards running way too hot for good longevity, and in some cases has them into thermal throttling even at STOCK clocks if they are in a room that isn't 70 F or lower.
DO NOT TRUST THE FACTORY FAN PROFILE, it WILL cause issues as it is not designed to handle "constant high load" applications such as folding.
 
I have yet to have an EVGA fan wear out, but I do tend to run "efficient" settings and don't normally need more than about 65% fanspeed to keep the cards reasonably cool and the oldest of my EVGA cards hasn't hit 2 years old quite yet.
Fan wear is VERY non-linear - 100% fan speed causes a LOT more wear than 80%, and 80% causes a LOT more than 60% - probably more like double the wear for each 20% increase as a ballpark SWAG.
 
Just don't get me started on the CHEAP JUNK fans used by Gigabyte on their non-Aorus cards - and recently on a few of their AORUS models.
 

yea, GPU boost starts to throttle down once temps get into the 80's. and with winter dropping temps about to go away i would keep an eye on the temps again over the next month.

Great write-up, Chris.  Thanks for taking the time and for replying to everyone who requested help; it's still very useful in 2019.

Duuude, I been searching all over the internet for weeks looking for a guide to do exactly what you posted.  As much time as I spend on these forums I have no clue how I missed it.  At any rate, thank you for taking the time to write such a detailed guide.  Yes, even today it's quite useful.  Thanks again