|
|
 |
|
|
Pages: 1
Windows XP ACPI Tweak Guide
(Click here to view the original thread with full colors/images)
Posted by: Ion Silverbolt
The guys over at Tweaker's Asylum has put up a guide on the effects ACPI can have on gaming, and how and what to do about disabling it if you want to try it. It might be something worth trying out if you want to attempt squeezing extra FPS out of your games. It especially seems useful for P4 performance.
Under ACPI I did very well on benchmarks on a Intel Pentium 4 2.0GHz and a GeForce3. I converted my Computer over to the Standard PC setup and rebooted only to see Windows XP find every piece of hardware on my machine. I installed the Intel 800 Chipset drivers for XP and NVIDIA GeForce3 drivers. I noticed the machine seem to be a bit zippier but how would it bench? As you will see, my benches showed a pretty big improvement.
Why would this be you ask?
Well, a theory could be that Windows XP ACPI has code for P4 to take advantage of Intel’s heat control options and could be lowering the power thus letting the P4 cool down. Turning this feature off disables the OS support for the cooling and runs at peak all the time improving processor performance.
Read the whole guide from here.
Posted by: Ion Silverbolt
Looks like a useful way to squeeze more out of a P4 processor. Only problem is, if your CPU fan fails or your heatsink falls off, it will most likely burn up the CPU instead of being protected.
So if the .003% chance of this happening scares you, do not disable ACPI. 
Posted by: Null Actor
ACPI has nothing to do with heat monitoring of processors, those guys are on crack.
The speed up probably comes from not having to delegate out interrupt requests through software (since everything shares a few IRQs and windows has to forward the requests to the appropriate hardware).
Posted by: Ion Silverbolt
I wasn't 100% sure. If the standard PC setting ignores BIOS calls, then I figured it might have somekind of effect on the heat monitoring capabilities of the P4 chip. It might have somekind of effect on the CPU-idle function even if the Heat monitoring of the chips still work.
I'm guessing the increase was a lot better with a P4 because not having to squabble for IRQ resources helped reduce haggling time with the memory controller. Thus, it didn't have to deal with Rambus latencies as much.
|
|
|
|
|
