Same computer, same dust - proof of overheating in current client
|Opened on:||Friday, March 21, 2014 - 18:02|
|Last modified:||Wednesday, April 2, 2014 - 19:39|
First - the good news for all the players who are overheating (and possibly the devs to track the issue down) - we aren't crazy, and I can prove it now.
Last night I checked my client usage and fan speed. I had two puzzles open, 861 & 862, and was running the same script on both - a Quaking Rebuild. My CPU fan was running at 2300+ rpm, and both clients were using 165-180% each (quad core). I closed the clients and rebooted the computer.
I keep old clients, and with karstenw's comments in mind - that this started back in early September - I opened a couple clients from last summer and selected a revisited and a second stage denovo to test, similar to what I was working on. That build is 20130720-4d4188ea95-osx_x86-devprev.
All tests were done with maximized clients, Gui showing and protein on screen.
My fan speed before opening the clients was 938. Opening the 2 clients and running the same script as above, Quaking Rebuild, brought the speed to about 1300. I opened another client and same script - speed 1400-1500.
A fourth client running the script - fan now at 2100. It took the 5th client to push the fan speed over 2300. 2300 is when I can hear the fan, and I try to stay under that speed. At each step, each client was using less than 100%.
I rebooted my computer and opened 2 current, Main clients and ran the same scripts. The fan jumped to 1700+ immediately. Opening a third client and running the script pushed the fan well over 2300 rpm.
So there is no question that at least for me, on my Mac, months ago I could run 4 clients with no heating problem, but anything over 2 clients now is a problem.
Same computer, same dust.
However… there is also no question that my bad habits from the past years when I could run the four clients are also contributing to the issue (my own, at least). After closing the third client, I left that script running overnight on both clients (minimized, protein off screen, Hide GUI) and 12 hours later, the fan was still reasonable at around 1300.
Maximizing them and showing the protein the fan is now at 2200+, both using around 140%. My lesson here? I had a uptime of 4 days before I started this process. Although I had closed and reopened clients during that time, apparently there was a memory build up that I was not flushing. This could be another indication for the devs, or just normal operation.
So I would recommend to all that you reboot your comps more often, and when finished with a script, save and close the client. Also - no question that especially now, clients must be minimized with the protein dragged offscreen and Hide GUI checked. This will be especially hard for hand folders, and for hand folding puzzles, but keep it in mind while we are waiting for the devs to fix the bug at their end.
Much a pain, but that may help until the devs can nail down the problem at their end.