Tom's Guide | Tom's Hardware | Tom's Games
![]() |
![]() |
![]() |
It's part of a program called "PC Alert 4" for MSI motherboards and is supposed to lower the temp of your processor. I figured no one would know what the hell i was talking about. =o)

Hi
The Cooler XP feature sends a kind of a Halt signal to the CPU, which can cool the CPU down anywhere from 5c to 15c, when Cooler XP is running you can still play games and do what ever you want to do, the program will not work if your using the CPU, it will only work when the CPU is idle
Not sure the exact way it works but it's an exelent progrmam for cooling the CPU
Zero Cool

CPUidle is software (download.com has it) that does the same thing. My t-bird runs at 51C under load and 47c idle, but with cpuidle it stays at 35C most of the time, going up to 51C during games or videos - these cpu halt programs can't help you when the cpu isn't halted.

How does it detirmine when to ingage it's "halting" process. My cpu is never at 100% idle. It usually bounces around from 95-98% when it isn't in use. This is because of anti-viruse softrware and the like.

When your cpu says its at 98% that means that over the previous 500 milliseconds it was busy for 10 of them. Without cpuidle your motherboard sent full power to the cpu for all of those 500 milliseconds, causing it to burn off heat like a light bulb with the power on high. With cpuidle your motherboard will send full power to the cpu for only 10 or 20 milliseconds out of these 500, allowing it to cool off. The result is that the cpu does not have all this constant amperage to burn off and cools down due to these breaks in the action.
Best to try it and see how it works. Here is my conclusion.
1) At full load cpuidle appears to act like its not there, since it never sends the HLT instruction to the cpu and the cpu runs at normal max load temp.
2) During browsing or editing or really anything but gaming and watching videos, the cpu has so many breaks that the temp drops about 15C.
3) During benchmarks, the start and stop delays caused by cpuidle cause a reduction in benchmark results of 2%-10%, something that bugs me but that I would not perceive had it not been for the becnhmark score. Interestingly, sequential block disk writes and memory writes seem to be the problem. My guess is that without cpuidle, these events have a huge amount of handshaking between the i/o and the cpu, such that the cpu is halted and restarted by cpuidle so many times a second that it causes an overall slowdown on that portion of the benchmark. If I try to forget about the benchmark effect, cpuidle produces no perceivable (to me) effect on performance.
4) Athlons act differently than other processors so CPUCool and Rain and Waterfall produce only a 3C lowering of the temp. Cpuidle does something else (they call it max athlon cooling) that makes it work with my t-bird way better than these other programs.

Without VCool (same thing than CPUIdle), my Athlon tbird 1,33ghz o/c to 1,43ghz run at 46C at idle. After being idle for 10 minutes, with VCool loaded, my temp drop to 25C, a 21C difference ! It's very COOL !

... because with cpuidle the cpu is halted instead of being allowed to idle. Windows determines processor usage by starting with the number 100% and then subtracting idle time divided by elapsed time. Since the cpu is rarely idle with cpuidle processor usage is reported as 100% much of the time. Windows would be better off by starting with 100% and then subtracting {idle time + halted time} divided by elapsed time.

![]() |
![]() |
![]() |

This post is quite old and has been locked from receiving new replies. Please create a new posting instead.
| Ads by Google |