View Single Post
23 Sep 2009  
H2SO4

Win7x64
 
 

Quote   Quote: Originally Posted by tmp View Post
From reading the posts in the forum fendytan, maybe it is a chipset problem. I've got Geforce 7150/nforce 630 combo integrated graphics board for laptops and as far as I can tell, have the latest drivers. Maybe nvidia is the place to direct this issue?

Fendytan-what chipset are you using? You said yours was nvidia too right?
Based on the latest info, I think you'd do well to update the chipset drivers and flash the BIOS. Given some of the things you mentioned and the fact that it happens on a clean build of Win7, it looks likely that there's an incompatibility between the OS and your current chipset/disk controllers which cannot be rectified by tweaking settings.

Interrupts are a way for a hardware device to get the OS's attention in response to some sort of time-critical event. For example, when the user presses a key, the keyboard issues an interrupt which makes the OS stop whatever it was doing for a (very) short while, deal with the implications of the keystroke, and then resume its earlier activity. Same with disks. When the disk controller has copied another unit-worth of data into a designated location for the OS to move, it issues an interrupt to let the OS know that there's something it needs to do. A large file copy is therefore a series of a squillion interrupts as each small chunk is copied by the HDD/controller into a buffer, then moved elsewhere by the OS.

One of the interesting things about interrupts is that they happen in what's called an "arbitrary thread context", which is a fancy way of saying that the OS temporarily hijacks whatever process/thread is running at the time to deal with the interrupt. In other words, if the HDD/controller happens to interrupt while calc.exe is running, that small chunk of processor utilisation will be "charged" to the Calculator, even though it had nothing to do with the file copy. Win7 is a bit better in this regard because at least it tells you that the processor utilisation stems from a high rate of interrupts, but because of the "arbitrary" nature of the process context there's no process name associated with that activity in Task Manager, as you've noticed.

When a hardware device goes a bit bonkers, sometimes it starts generating interrupts at such a high rate that it effectively pesters the OS to the point where it cannot do much else. It may look like some hapless process is consuming that processor time, but end-tasking the process doesn't help because the interrupt utilisation simply gets attributed to the next one to come along. I think that's exactly what's happening in your case, except the HDD controller may not be ga-ga... it's just got a fundamental low-level incompatibility with Win7, for reasons that I can only guess at.

If the machine has firewire (1394) or eSATA ports, it would be very interesting to see what happens if you plug in external drives of those varieties, borrowed if necessary. My guess is that the problem probably wouldn't occur, given the somewhat different mix of hardware and drivers involved.

I am not aware of any HAL updates for RTM Win7 at this point, but I'm not well informed on that topic. At this stage, my suggestion would be to go for every single update you can find (from the manufacturer/developer, not from a 3rd-party), and to cross your fingers that will do the trick. The fact that it also happens in safe mode, with multiple completely "clean" OS installations, and with multiple drives, pretty much rules out something that could be fixed with a tick in a tickbox.

Lastly, you've tried all of the USB ports on the machine and they all do this, right?
My System SpecsSystem Spec