[KanagiF]

(tl;dr version is here http://www.tomshardw...nojs=0#t2643974 which ironically is even longer).

Built a new comp on Saturday; When I first built it it ran great, nothing out of ordinary. That night I noticed Core 1 was always 60+% usage no matter what. googled it, it let me to find out it's Hardware Interrupts which I read A LOT of threads and tries hundreds of different things about. ATM I am downloading a different Windows CD to see if maybe my Windows is just jacked up, but if that isn't the case I'm not sure what to do.

Brought the PC to work with me today to work on it and for the first 3 hours, the problem was actually completely gone. At some point in between me being in the office, and in the room where it is, it came back.

I've run XPerf and I have the graph but in all honesty I have no idea what I'm looking for.

I've been using LatencyMon to track it, and earlier it wasn't coming back with any errors (For once) until I ran something then the same 3 as usual came up. USBbus.sys, atapi.sys (or Iastor.sys depending is I have Windows or Intel's IDE drivers), and HDAudBus.sys .

I'm -really- hoping it isn't hardware, especially since it's gone away a couple times.

I've tried switching around all hardware, formatting on different drives, etc etc. Some of these drivers though I'm not sure how to "fix". For instance Uninstalling USB drivers just makes me unable to do anything, (I tried using a PS2 mouse but still didn't fix my issue).

Is there anyone who can read XPerf sheets or can give me guidance? If it IS hardware (I assume it'd be the mobo) I can hopefully take it back and exchange it tonight. I was just really excited about using it, and so far I've just been staying up until 2/3AM every night working on it.

Thanks ~ Slow at work so I'll be checking back/replying as soon as as answer/tip is provided.

[MagicAndre1981]

yes, send me the xperf generated ETL file.

[MagicAndre1981]

Quote

DPC Info

--------------------------
CPU Usage Summing By Module For the Whole Trace

Elapsed Time, > 256 usecs AND <= 512 usecs, 259, or 1.46%
Elapsed Time, > 512 usecs AND <= 1024 usecs, 179, or 1.01%
Elapsed Time, > 1024 usecs AND <= 2048 usecs, 241, or 1.36%
Elapsed Time, > 2048 usecs AND <= 4096 usecs, 126, or 0.71%
Elapsed Time, > 4096 usecs AND <= 8192 usecs, 1, or 0.01%
Total, 17699

Total = 5653 for module ndis.sys
Elapsed Time, > 1024 usecs AND <= 2048 usecs, 231, or 4.09%
Elapsed Time, > 2048 usecs AND <= 4096 usecs, 126, or 2.23%
Elapsed Time, > 4096 usecs AND <= 8192 usecs, 1, or 0.02%
Total, 5653

Total = 467 for module nvlddmkm.sys
Elapsed Time, > 512 usecs AND <= 1024 usecs, 37, or 7.92%
Elapsed Time, > 1024 usecs AND <= 2048 usecs, 10, or 2.14%
Total,

so you have NDIS and nvlddmkm.sys issues.

The NDIs issue seems to be caused by WinRAR. What are you doing with WinRAR?

Can you run msinfo32.exe and look if the NVIDIA GeForce GTX 560 Ti shars the same IRQ like the network card?

[KanagiF]

WinRar I actually JUST installed when you needed me to zip the files for you. My comp is mostly a fresh install.

As for IRQ: My video card is on 16, which is also:


IRQ 16 Creative SB X-Fi OK
IRQ 16 Intel® 7 Series/C216 Chipset Family PCI Express Root Port 1 - 1E10 OK
IRQ 16 Intel® 7 Series/C216 Chipset Family PCI Express Root Port 5 - 1E18 OK
IRQ 16 Intel® 7 Series/C216 Chipset Family USB Enhanced Host Controller - 1E2D OK
IRQ 16 Intel® Management Engine Interface OK
IRQ 16 NVIDIA GeForce GTX 560 Ti OK
IRQ 16 Xeon® processor E3-1200 v2/3rd Gen Core processor PCI Express Root Port - 0151 OK


Sorry while doing this my mouse stopped working (Whenever that happened USBbus.sys or whatever shoots up.

[MagicAndre1981]

Connect the mouse to different USB port which doesn't share the IRQ with the nVIDIA. If you have 2 different PCIe x16 slots plug to nVIDIA to the second slot.

[KanagiF]

MagicAndre1981, on 12 September 2012 - 01:31 PM, said:

Connect the mouse to different USB port which doesn't share the IRQ with the nVIDIA. If you have 2 different PCIe x16 slots plug to nVIDIA to the second slot.

I do have a second PCI port. Let me try that and I'll get back with you. Do you need another file after-wards, or see if the interrupts go away and we'll go from there?

[MagicAndre1981]

yes, make a new trace after using the second PCIe x16 slot.

[KanagiF]

https://skydrive.liv...DD1A0D8BCE1!124 is the new trace

[MagicAndre1981]

it is now better:

Quote

DPC Info

--------------------------
CPU Usage Summing By Module For the Whole Trace

Total = 10268
Elapsed Time, > 0 usecs AND <= 1 usecs, 3132, or 30.50%
Elapsed Time, > 1 usecs AND <= 2 usecs, 807, or 7.86%
Elapsed Time, > 2 usecs AND <= 4 usecs, 341, or 3.32%
Elapsed Time, > 4 usecs AND <= 8 usecs, 1505, or 14.66%
Elapsed Time, > 8 usecs AND <= 16 usecs, 1377, or 13.41%
Elapsed Time, > 16 usecs AND <= 32 usecs, 1528, or 14.88%
Elapsed Time, > 32 usecs AND <= 64 usecs, 946, or 9.21%
Elapsed Time, > 64 usecs AND <= 128 usecs, 409, or 3.98%
Elapsed Time, > 128 usecs AND <= 256 usecs, 119, or 1.16%
Elapsed Time, > 256 usecs AND <= 512 usecs, 34, or 0.33%
Elapsed Time, > 512 usecs AND <= 1024 usecs, 65, or 0.63%
Elapsed Time, > 1024 usecs AND <= 2048 usecs, 5, or 0.05%
Total, 10268

now there are 5 spikes over 1ms (1024µs)

Quote

Total = 2099 for module ndis.sys
Elapsed Time, > 512 usecs AND <= 1024 usecs, 1, or 0.05%
Elapsed Time, > 1024 usecs AND <= 2048 usecs, 1, or 0.05%
Total,

the nvlddmkm.sys usage seems to be caused by the Sidebar. Which gadgets do you use?

[KanagiF]

Haha, ironically Core Temp and GPU Temp. So I can watch the core usage.

I still have some conflict IRQs. Dunno if that matters.


Sound card is on the same as the USB ones (16, where the video card used to be.)

and then

IRQ 18 Intel® 7 Series/C216 Chipset Family PCI Express Root Port 7 - 1E1C OK
IRQ 18 D-Link DWA-556 Xtreme N PCIe Desktop Adapter OK
IRQ 18 High Definition Audio Controller OK

makes me wonder, only because I had an HDAudBus.sys file going berserk on me, and wasn't sure if that could be why.

This post has been edited by KanagiF: 12 September 2012 - 02:12 PM

[MagicAndre1981]

Disable them and test again.

An users also had GPU issues and the casue was power saving options for PCIe. Go to the power saving options and disable all PCIe power saving options. Maybe this helps, too.

[KanagiF]

https://skydrive.liv...DD1A0D8BCE1!125

New test with Gadgets disabled, Sidebar closed, and power savings disabled for PCI-E.

[MagicAndre1981]

still nVidia driver spikes:

Quote

--------------------------
DPC Info

--------------------------
CPU Usage Summing By Module For the Whole Trace


Total = 251 for module ctoss2k.sys
Elapsed Time, > 256 usecs AND <= 512 usecs, 125, or 49.80%
Elapsed Time, > 512 usecs AND <= 1024 usecs, 105, or 41.83%
Total, 251

Total = 6 for module iusb3xhc.sys
Elapsed Time, > 1024 usecs AND <= 2048 usecs, 3, or 1.56%
Elapsed Time, > 2048 usecs AND <= 4096 usecs, 18, or 9.38%
Elapsed Time, > 4096 usecs AND <= 8192 usecs, 2, or 1.04%
Total, 192
1

Total = 825 for module nvlddmkm.sys
Elapsed Time, > 512 usecs AND <= 1024 usecs, 124, or 15.03%
Elapsed Time, > 1024 usecs AND <= 2048 usecs, 9, or 1.09%
Total, 825

Now i can see that Chrome causes the nVIDIA driver issues. Try to disable HW acceleration in Chrome.

Now you also have USB 3.0 driver issues. If you don't need USB 3.0 disable the controller. Also can you plug the D-Link DWA-556 Xtreme N PCIe Desktop Adapter into a different PCIe port so that it no longer shares an IRQ wit hthe sound card (ctoss2k.sys - Creative SB X-Fi)?

[KanagiF]

I'll look at Chrome, I didn't know it had HW acceleration. As for the sound card/wireless, both are PCI-E and I only have 2 PCI-E slots. but I can take the sound card out until we see for sure what the issue is, and I'll send you the new-new report. Gimme just a second.

[KanagiF]

Just an update: I played with it for about 4 more hours (About 80 hours spent on it total.) and I got IRQ's down to almost none (Except of course like Intel PCI on the same as video card, etc) and still had 30-40% CPU Interrupt. I rage quit and finally removed the motherboard and returned it. Well, exchanged it because I assumed it was a fault board. Just got comp all set back up and installed everything and interrupts haven't really passed 1%, or even less. So I guess I just had a faulty board. Thanks for all your help, Andre.

[MagicAndre1981]

ok, nice to see that a new mobo fixed it