Title: XTweaks : Cool when Idle and Windows 8 Post by: PeterSt on January 19, 2013, 04:54:59 pm In Windows 8 the Cool when Idle parameter can best be set to 1 (Cool) because otherwise the usage of all CPU cores jump to a certain percentage (depending on the Balanced Load setting). Whether it is really true that something makes the cpu cores busy is unclear, but it looks to be really so.
If you are comfortable with it anyway, you can use it (not Cool) if you want. Peter Title: Re: XTweaks : Cool when Idle and Windows 8 Post by: Scroobius on February 07, 2013, 03:03:22 pm I agree some strange things happening with CWI set to 0 which is where mine is now. But I find similar behaviour in W8 as it was with W7.
I have a Zalman Reserator a nice tall cooling tower (I couldn't let Mani and Nick have all the fun could I?). And I always like to "feel" it at odd times (bear with me!!) to see how warm it gets. And it can certainly gets warm with CWI=0 BUT not all the time it seems to go in cycles sometimes warm sometimes only luke warm. The temperature of the cores is relatively constant at 44-46 for 5 cores and 48-52 core six. While temperature of the cores is OK now I would not want to leave it in summer with CWI=0. P Title: Re: XTweaks : Cool when Idle and Windows 8 Post by: PeterSt on October 26, 2013, 11:58:08 am All,
I finally found that the Cool when Idle = 0 (= not cool) really drives the CPU mad in Windows 8 (not so in Windows 7). A bit like Paul said, I can hardly see it on the temperature rising (and if that rises it is one degree Celcius only). However : I now found that the draw on the mains is significantly higher. A whopping 24% more (measured on the usage of the whole PC, in my case the "XXHighEnd PC" with 6core 3930. Amperage used about the same story (19%). This means that the ~37% usage TaskManager shows is somehow for real (and an OS bug). Do notice that not all "monitoring" programs show this. But if they do not, something will have been tweaked. Also think about the "latency" results on the Latency Checker program, which also had to be adjusted for W8. So here too, we must be careful whether we see the real thing. In this case the implied higher draw tells all. Peter |