I posted this question a while back here and received several suggestions on how to proceed, but now I'm stuck again.
CPU usage jumps to 100% -> WHEN <- machine is idle (not the idle process)
Similar question here:
Task Manager shows 100% CPU utilization, but nothing in process list does.
I downloaded and installed Process Explorer, as Greg recommended and it did make things a bit clearer, but I'm not any further along in fixing it. The process that is hogging all the CPU is simply listed as 'Interrupts', with no further details. I've Googled that and found lots of people writing about it, suggesting uninstalling disk drives, changing I/O protocol and such, but nothing of what I have tried has had much effect. Sometimes the machine is a little better behaved, sometimes not, but WinAmp reliably invokes this behavior.
Once again, this is relatively new behavior, so I don't believe that the machine simply isn't up to the job. And most of the time, it plays normally, at very low CPU usage, as I would expect from a low-stress activity like playing MP3 files. But the periodic 100% spikes remain and I'm at my wit's end trying to figure out what's causing it.
Okay, if anyone happens to be on, I'm at the machine now.
Hyppy, you were right, the DMA/PIO settings are in the controller properies. My work machine has SATA drive only, but this one has IDE drives and the properties are as you said.
The properties listed are as follows:
Primary channel IDE
Device 0 - transfer regime: DMA (if available); actual transfer regime: Regime Ultra DMA 2
Device 1 - transfer regime: DMA (if available); actual transfer regime: not available
Primary channel IDE
Device 0 - transfer regime: DMA (if available); actual transfer regime: not available
Device 1 - transfer regime: DMA (if available); actual transfer regime: Regime PIO
Secondary channel IDE
Device 0 - transfer regime: DMA (if available); actual transfer regime: not available
Device 1 - transfer regime: DMA (if available); actual transfer regime: not available
Secondary channel IDE
Device 0 - transfer regime: DMA (if available); actual transfer regime: not available
Device 1 - transfer regime: DMA (if available); actual transfer regime: not available
I'm going to restart the machine and look in the BIOS to see if there's anything there that will shed some light on this.
Well, tried a restart, after setting IDE stuff in BIOS to Auto, that had some specific types of transfer modes set. No luck, device 1 in the second Primary channel is stil in PIO mode. I found some very interesting looking stuff here:
but the owner of the teashop was starting to get cranky (it was 1 am here) so I gave up for the night. But I'm going to keep posting my progress here, as long as the moderators let me keep this thread open. It's a monumental pain in the fundament, the most aggravating troubleshooting job I've encountered in quite a while, but I don't intend to give up. I'm learning things, and just maybe it'll also help someone else down the road.
DMA and StudioHack - thanks for tidying up my stuff here. I'm pretty new on this site and don't have all that good a grip on how it works yet. I was pretty active in some MS newsgroups a while back, but this is quite different - appreciate the help.
SUCCESS!!!!! At least it appears so, at this point. I examined and ran the script offered on this site:
http://winhlp.com/node/10
then restarted the machine and it seems to have been the final straw that did the trick. The two active channels/devices have both reset to Regime Ultra DMA 5, the stuttering sound and 100% CPU usage are gone and the entire machine reacts considerably faster to everything, including start-up. It's almost as if it has risen from the dead. Many thanks to everyone that helped with suggestions and explanations - I believe that this can be considered solved and the thread may be closed.
