The only one I can recall is for nForce2. It fixed up C0/C1/C2 states for that chipset + Athlon XP.
http://www.softpedia.com/get/Tweak/System-Twe … ing-Patch.shtml
`

Some mainboards have a stpgnt/halt option in the BIOS setup. If not then here you can find CoolON 2.1.0.

Thanks gerwin! Unfortunately, my BIOS does not allow me to set these states.

Are you using this on your K7T Turbo? Does it seem to work?

Yes, my K7T Turbo with the same XP 1800. It seems to be working well, peak temp before was 56 and now is 53, and more commonly stays in low 40s during high load. And idle temp gets down to 30. I didn't do a strict before and after test, so maybe I will, but I'm satisfied with those numbers.

May I also ask what OS you are using? With my 8KTA3 and Mobile Barton, I have successfully enabled the halt states in a number of ways, with Cool-On, wpcredit, and such. Sometimes I can see that there is a definitive, marked improvement, and then after an undetermined amount of time it freezes. The other times it simply freezes right away! Based on posts that I have read, there seemed to be two theories as to the cause. One said that it had something to do with the OS, with people stating that they didn't have a problem in 98, and that NT based OS's did. I also saw another one where there was speculation that it was a motherboard issue, specifically the voltage regulators not being able to handle the huge swings in power draw as the processor switches in and out of the HLT / Bus disconnect modes. I really wanted to get this working since my machine is left on... well pretty much constantly. In my current configuration, my processor is under-volted from the original Athlon XP spec of 1.65 to 1.475. During idle it usually hovers around 41-42, and even while under full load it rarely tops 50. However, after one particularly long run with the cooling on it crept all the way down to 34.

I'm using Win2k SP4. Just to be complete, my CPU fan is a Whisperrock IV, and the PSU has a fan directly above the CPU which draws air off it. Those are the only two fans in the system. Ambient temp makes a big difference too - nighttime is when I can get down to 30 at idle, daytime is higher, but I don't use it much during the day. I haven't noticed any stability issues at all so far.

I think I'm going to do an extended test without either state enabled to see what happens. Of course there will be a report here.

It is apparent that most mobo manufacturers didn't care to implement AMD's new idle states back with K7. So it is entirely possible that the motherboards/chipsets are not designed to deal with the effects of forcing them enabled.

Modern boards usually have more beefy power regulation than older boards. 2-3 phase designs with cheap components are what I remember. I imagine that lots of those boards probably used the defective capacitors of the day, too. Newer designs do take into account the big changes in power use when coming in and out of CnQ/EIST/C1E/etc. The change in power use with a modern CPU is also much more dramatic however than with a K7 that remains at the same clock & voltage all the time. If the mobo power regulation can't deal with the change in load, the CPU can get overvolted or undervolted depending whether it is going into idle or coming out of idle.

You can read a bit of info about modern power regulation on mobos here (mainly Intel designs):
http://www.thetechrepository.com/showthread.php?t=126

I suppose it is a mixed bag, since some motherboards appear to work ok and others don't. I really like my Athlon platform despite the few shortcomings it has. I wanted to feel around and see if I could identify motherboards that supports the cooling and multiplier adjustments, at which point I'll likely get it re-capped and spiffy-fied for ultimate function and longevity. I haven't really messed with mine too much since it isn't idiot-proof / rock stable yet, and since it is my main machine I try to keep it in a mostly working state at all times. Pretty soon I'm getting a neat Pentium D motherboard that will be my primary, and then I'll have the means to continue sorting out issues and not compromise as much.

Well, I did a rough test. Without either state enabled, idle temp was 56. With stop grant only it was 51. When halt was then enabled it dropped to 38 in about 60 seconds and to 36 a few minutes later (this was during mid-afternoon after one hour of idling).

So this motherboard seems to have no trouble with these states. However, in the past I have found that it hates having its multiplier or FSB changed.

I have read that there is a register you need to set on the KT133A in order for it to catch the FDIV changes. It is register 55 Bit 2 - FID Command Detect, 0=disable, 1=enable. This however has not worked for me. You could try that and give the multiplier adjustment a go. I think you need to be using an unlocked chip though...

Hm, well, I'm not really too interested in overclocking anyway. Thanks for the info though.