WattMan Crashes & Revert to "Balanced" Settings

Discussion created by heartsbane on Nov 17, 2017

I am running applications on my system that use 100% activity of my Radeon RX Vega 64 for several hours (unmonitored). With the default Wattman settings ("Balanced") the GPU rises to 80+ C in the first 60 seconds. Even at this temp, I noticed the fans only spin at ~1,200 rpm (25% of max). I decided to tweak the fan/temp settings in WattMan in order to improve the operating temp of the GPU while it is operating at 100%.


I changed the setting to "Custom" and the only settings I changed were: Fan Speed (2400 min, 4800 target), Temp (85 max, 60 target), and Power Limit (-11%). So my custom settings have a lower power limit and significantly more aggressive fan settings than "Balanced." When I apply these settings, they have the desired effect of lowering my GPU temp to around 55-60 while operating at 100% (fans are spinning at ~4300 rpm). So I walk away from the PC thinking everything is great.


15 minutes or so later, I come back to GPU that is 85+ C, and fans that are spinning at < 1,500, and WattMan has reverted back to the "Balanced" settings. This has happened several times, and so I watched it and noticed that WattMan is experiencing sporadic crashes and reverting back to the "Balanced" settings, which then cause the temperature to quickly skyrocket into the 80's since the "Balanced" settings don't allow the fans to spin up aggressively.


So I have several issues/questions about WattMan, and I'm hoping someone out there can provide me with some answers:

  • Is there a way to modify the "Balanced" fan/temp settings to make them more aggressive than the default? If not, can I prevent WattMan from defaulting back to these less aggressive settings on a crash (and therefore exasperating the issue)? In my case, this behavior is extremely frustrating and potentially damaging to my GPU!
  • What's up with WattMan crashing so much? I have tried to determine a specific cause, but it seems to happen sporadically. Is there a log somewhere that indicates the cause of the crash or any diagnostic info? How can I determine the root cause (and hopefully fix it)?
  • I tried creating app-specific settings for the application I'm running, hoping to override the global settings. However, the app-specific settings don't appear to work at all, even when I launch the program from WattMan, the fan/temp settings are unaffected. I clicked on "Check Compatibility" and my app-specific settings don't have a green check mark by them (other profiles I see are mixed, some have green check, some don't). Why have this feature if it doesn't work?!?
  • Why are the default "Balanced" settings so crappy? If my GPU starts to climb over 75, the fans should be kicking in AGGRESSIVELY to control the temp. The default settings for both fan speed and temp max/target wait too long and do far to little to control the temp of my card.


Overall, I'm very frustrated with not being able to control my fan/temp curves and having WattMan revert to default settings that put my GPU at risk of being damaged.




System Info:

- Custom PC, Windows 10 Pro (64 bit), 32 GB RAM

- CPU: AMD Ryzen 7 1800X

- GPU 1: nVidia GeForce GTX 1070

- GPU 2: AMD Radeon RX Vega 64 (8GB)

- Radeon Software Version: 17.11.1 (Settings Version: 2017.1102.1456.25052)