timtoflly wrote:...What does this line mean? 17:02:17:WU01:FS01:0x17:Temperature control disabled. Requirements: single Nvidia GPU, tmax must be < 110 and twait >= 900
This is a feature present in FahCore_17 version 0.0.52 and higher which if configured, will allow the GPU to pause folding for a time period once it reaches the configured temperature. By default, it is disabled but you can enable it if you want to.
I didn't realise it was "active but disabled", I thought it was a placeholder for a future version.
A few questions:
Is it simply a matter of entering suitable values for tmax and twait as slot options for the appropriate GPU, or is an enabling entry also required (eg "enable-temp-control" = "true")?
I assume the values represent ºC and milliseconds?
Does it simply pause processing or does it also write out a checkpoint file?
Might be a way to get my 780Ti back under Linux instead of having to use a Windows app in order the control the temperature, though the on/off operation might cause too much PPD loss compared to more gradual control.
Have a look here: viewtopic.php?nomobile=1&f=89&t=25075
With cool-bit you can enable the fan control also under Linux. I use it for my GTX 780s and 660ti. I just keep the fans on 75% - 85%, depend on ambient.
With nvidia-smi you also easy can read the temps from a script if required. No need for windows here.but I also don't bother as they anyway run 24/7 and need constant air flow.
If you want real overclocking, Windows might be better. No experience with that though.
Thanks, but I tried using coolbits with Linux- with the degree of factory overclock on the card (base is 1020MHz) I need to run the fans at 90+% to keep it stable in the evening when the sun hits the windows (it runs 24/7 mostly unattended), and even then it's marginal on some WUs. Not good for fan longevity
I want it to look after itself, not have to keep changing settings whenever the weather changes- don't forget I'm in the UK and that could be a full-time job
With Windows I can set a mildly aggressive fan profile (but one that won't wreck the fans), limit the temperature to 65ºC and it's still stable even with a further 25MHz of overclock… the app only reduces the clock/core volts when necessary so whenever it can it runs flat out at 1188MHz. If I can get somewhere near that with Linux, so much the better
This is the argument format:
<extra-core-args v='-tmax=TMAX -twait=TIMEINSECONDS'/>
Where
TMAX is the GPU temperature and the value should be less than 110 (degree Celsius)
TIMEINSECONDS is the time that the GPU will pause once the TMAX has been reached. Minimum waiting period is 900 seconds (15 minutes) since frequent temperature variation would stress out the GPU (thermal stress is something that should be avoided).
You can do that by following this method:
1) Open up Advanced Control (AKA FAHControl)
2) Click on Configure
3) Select the Expert Tab
4) Under Extra core options add this (change the values of tmax and twait to your choice): -tmax=78 -twait=900
5) Click OK
6) Click Save
I believe that the next time FahCore_17 starts, the arguments will be passed to it and the message would be something like this:
23:17:06:WU00:FS01:0x17:core17: single GPU Temperature Control enabled, tmax: 78 twait: 900
ETA:
Now ↞ Very Soon ↔ Soon ↔ Soon-ish ↔ Not Soon ↠ End Of Time
Hardware configuration: Intel i7-4770K @ 4.5 GHz, 16 GB DDR3-2133 Corsair Vengence (black/red), EVGA GTX 760 @ 1200 MHz, on an Asus Maximus VI Hero MB (black/red), in a blacked out Antec P280 Tower, with a Xigmatek Night Hawk (black) HSF, Seasonic 760w Platinum (black case, sleeves, wires), 4 SilenX 120mm Case fans with silicon fan gaskets and silicon mounts (all black), a 512GB Samsung SSD (black), and a 2TB Black Western Digital HD (silver/black).
Does it simply pause processing or does it also write out a checkpoint file?
Just like the PAUSE function and like controlling the GPU with IDLE, there are no special cirumstances for checkpoints. Processing is simply suspended to limit the temperature.
jrweiss wrote:Don't all the modern nVidia and AMD GPUs have auto-throttling for high temps built in?
I don't know enough to answer that properly, but if they do it is probably aimed at protecting the gpu against damage rather than maintaining the integrity of the data it may be processing.
My card introduces errors at high temperatures on folding without more control than anything built-in to the card or drivers provides.