![]() |
High CPU usage for hardware interrupts |
Post Reply
|
Page <1 4567> |
| Author | |
Ramin99
Newbie
Joined: 16 August 2008 Location: Canada Status: Offline Points: 4 |
Post Options
Thanks(0)
Quote Reply
Posted: 16 August 2008 at 3:37am |
|
The problem doesn't occur in safe mode, but I'll have to run safe mode for days to be sure.
Yes, I do have the latest drives. I checked/cleaned/reseated everything except the CPU, and I didn't do Chkdsk for the hard drive. I will do them tomorrow, then wait for the next spikes. |
|
![]() |
|
thebrammer
Newbie
Joined: 25 July 2009 Location: scotland Status: Offline Points: 2 |
Post Options
Thanks(0)
Quote Reply
Posted: 25 July 2009 at 8:51pm |
|
i dont know if this will be of any help but i too was having this problem of interrupts causing high cpu usage. have found that after re-installing my motherboard drivers the problem has gone away. abit i-45cv mobo. seems that the mobo was having trouble speaking to my hardware. ram etc.
might help to look to see if there are any updated drivers for your mobo. also found that this is only and issue when running windows. linux seems to know how to work it from the start. |
|
|
ribz
|
|
![]() |
|
thebrammer
Newbie
Joined: 25 July 2009 Location: scotland Status: Offline Points: 2 |
Post Options
Thanks(0)
Quote Reply
Posted: 30 July 2009 at 1:18pm |
|
additional. i have also found that since i have uprated the cooling on my cpu the problem also lessened. pent d925. 3.0 dual core. puts out major heat. dont know if this is of any interest but the pent d doesnt like to run higher than 60 deg c.
|
|
|
ribz
|
|
![]() |
|
espin
Newbie
Joined: 15 August 2009 Status: Offline Points: 1 |
Post Options
Thanks(0)
Quote Reply
Posted: 15 August 2009 at 11:12pm |
|
Thanks to this forum, as I fixed my problem.
Details of my problem and it fix. Problem was that HDD (in the laptop) had 2 bad sectors, one in pagefile, so windows when see errors on communication with HDD down-grading drivers (from Ultra DMA to PIO - please check if this apply to you), so I have fixed bad sectors (took 12 hours because PIO), next to fix the driver you need delete it (in my case IDE ATA/ATAPI controllers | Primary IDE Channel) and restart windows, so it will install Ultra DMA. Problem sorted.
|
|
![]() |
|
Kimo
Newbie
Joined: 09 October 2009 Location: Swiss Status: Offline Points: 1 |
Post Options
Thanks(0)
Quote Reply
Posted: 09 October 2009 at 11:40pm |
|
Hi all
Just registered to add my experience. My machine is 6 month old and from the start I had spontaneous BSODs, black/distorted screens, freezes, etc, mainly in 3d-applications, but not only. Very hard to nail down – I'm loving it! So I went for the VGA-card, then for overheating, to finally find out it was one of the RAM-sticks. After all the system was not stable anymore and I did a new install with updated drivers, software and all – and got 45% CPU-usage through interrupts. Following this thread and google I found the source: Marvell 61xx Raid Controller. I miss my old IDE-Burner now, but hey, CPU works! :) I'll look into BIOS-settings and (re)installing the (another) driver later. After all this hours/days/months I don't feel like tinkering anymore right now... In future, if any miracolous problems occour, I'll go straight for the RAM! Thank you! Asus P5Q-E Intel 8500 w/ Arctic-Cooling Freezer Extreme (this one rules!) OCZ 4x 2GB (PC-1066) (today a bit surplus, but I wanted to be futureproof and to have 4 sticks with the same specs. And Vista is not sad about it. ) Asus 4850 Matrix 2x WD 500 GB Raid 0 T-Balancer Fan Controll (for silence, no o/c) Vista 64 Home Premiun |
|
|
Who only knows linear thinking will sometimes skid out of the turn in our round world...
|
|
![]() |
|
SFDreamerX
Newbie
Joined: 24 October 2009 Location: United States Status: Offline Points: 1 |
Post Options
Thanks(0)
Quote Reply
Posted: 24 October 2009 at 9:02pm |
|
Dear Awesome Forum!
I found a solution for my 50+ percent "Hardware Interrupts" CPU utilization problem on another forum, which I wanted to share. The solution was to disable the IEEE 1394 device. The post that "saved me" mentioned disabling the "OHI Compliant IEEE 1394 Host Controller" in the device manager. I found that disabling the IEEE 1394 device in BIOS worked just as well. I have no idea what triggered this problem, but this solution was OK for me since I don't use the IEEE 1394 device (i.e. FireWire, i.LINK, Lynx).
This "Hardware Interrupts" problem just popped up within the last week, which was weird. Since this machine uses a SATA drive, the IDE PIO solution did not apply to me. Thank goodness for the quality of reporting offered by Process Explorer! It would be nice if Mark Russinovich had a Process Explorer plug-in or a separate utility that would take a snapshot (variable time period ms to secs?) of interrupts by IRQ. This would also offer one more parameter for providing insight into performance bottlenecks (i.e. a higher than normal interrupt activity from a specific device).
Regards, SFDreamerX
====== Related Hardware Info ========
Motherboard (purchase 1/2006)
(Sckt939)ASUS A8N-SLI nForce4 SLI Chipset SATA RAID Dual PCIE MB w/Gb-LAN,USB2.0,IEEE-1394,&7.1Audio [+34]
CPU: (939-pin) AMD ATHLON64 3200+ CPU w/ HyperTransport Technology
|
|
![]() |
|
molotov
Moderator Group
Joined: 04 October 2006 Status: Offline Points: 17506 |
Post Options
Thanks(0)
Quote Reply
Posted: 26 October 2009 at 3:20am |
|
Check out Kernrate or RATTV3, or for Vista and later xperf from the Windows Performance Toolkit.
|
|
|
Daily affirmation:
net helpmsg 4006 |
|
![]() |
|
irha
Newbie
Joined: 12 April 2010 Status: Offline Points: 1 |
Post Options
Thanks(0)
Quote Reply
Posted: 12 April 2010 at 4:08am |
|
Thanks for the suggestion of RATTV3, I was able to narrow down to my serial port (I think). I have been having trouble with one of the core's going 100% after the system comes back from standby or hibernate and this has been bugging for a long time, as the only solution is to reboot. After analyzing the RATTV3 log, I figured that serial had a very high count and so tried disabling it first with no luck. But when I reenabled it, the CPU magically went down to almost 0%. I don't know what actually is causing this, but since I am not really using the serial port, I can keep it disabled for now.
|
|
![]() |
|
okiro
Newbie
Joined: 18 October 2010 Status: Offline Points: 1 |
Post Options
Thanks(0)
Quote Reply
Posted: 18 October 2010 at 10:45pm |
|
Oh my... I had the same problem and I have the 100% solution IF you are using graphic card AGP ATI RADEON HD3650. As I understood on other forums this card combined with motherboard of NVidia chip and IRQ on ATI doesn't work well. I mean HDMI doesn't work which causes all trouble.
My problem log: (CP>Admin Tools>Event Viewer - perfect place to detect problems and errors) "IRQARB: ACPI BIOS does not contain an IRQ for the device in PCI slot 0, function 1." Solution: (if you are using ATI HD3650 and Nvidia chip mb) In 'Device Manager', expand 'System Devices', disable 'Microsoft UAA Bus Driver for High Definition Audio'. That's all. Well if you need HDMI you can use various hacks, but for them search google more.. :) BTW: I registered here just to help you people and I hope i will help somebody... |
|
![]() |
|
DetroitDave
Newbie
Joined: 29 December 2010 Location: MO Status: Offline Points: 1 |
Post Options
Thanks(0)
Quote Reply
Posted: 29 December 2010 at 9:31pm |
|
Finally! I found your site and this (totally simple) solution to my hardware interrupt
problem! XP had indeed reset my IDE/ATA controller settings from DMA to PIO. Removing the
driver and restarting the pc fixed it. Meanwhile, I had spent a week looking around for a
solution to my pc slowdown problems (Toshiba Satellite laptop with a
Hitachi hdd) until I almost accidentally ran across this thread on your
excellent forum.
Apparently this issue -- Windows resetting DMA transfer to PIO -- is fairly common (see: http://support.microsoft.com/default.aspx?scid=kb;en-us;817472, for instance) -- yet it's not the among the common solutions that show up when searching the internet tubes for "system slowdown" (or the like) issues. I don't know how many wild goose chases I followed -- (virus?, hard drive or other hardware issues? malware? etc., etc.) before discovering the simple solution discussed here. Many, many hours wasted. Sigh. Which reminds me to ask: If Windows, based on things like a few bad restarts in a row, more or less routinely resets the drive controller to use PIO instead of DMA, then why does it not give a large WARNING about what it's doing, why it's doing it, and mention that it will severely hamper system performance and maybe also mention, by the way, how to reverse the setting(s)? Maddening!! Finally, though I am not a total noob, casual suggestions to "delete drivers and let XP fix them on reinstall" can seem a bit disconcerting to the ordinary user. At the risk of redundancy (and boring the geekier reader to death) here, for the ungeek, is the full step by step way to check to see if XP has turned off your DMA mode and how to get it back if that's the case. This assumes a mostly vanilla setup with a standard IDE/ATA drive controller which is probably the case for most pc's that have this particular malady. Right click "My Computer." Choose "Properties." Select the "Hardware" tab. Select the "Device Manager" button. Expand (click the + sign) next to "IDE ATA.ATAPI controllers" Right click "Primary IDE Channel" Select "Properties" Select "Advanced Settings" ... here you will now see the ATA/IDE devices for your system. Typically you are concerned about your system hard drive which is called "Device 0" (zero). There are three boxes under Device 0: Box 1: Ensure that "Device Type" is set to Auto Detection Box 2: "Transfer Mode" should be set to "DMA if available" Box 3: should read "Ultra DMA Mode 5" (or something similar) If that is how it all reads, this is not your problem. But iff Box 3 reads "PIO" then XP has indeed reset your transfer mode to the slowest setting which will cause up to 100% kernel processing time for long periods of time to accomplish even simple operations.. So now you know what your main problem is! To fix this don't go anywhere! Stay in the same Device Manager "IDE Channel Property Settings" dialogue box. Select the "Driver" tab at the top of the dialogue box (next to "Advanced Settings"). Select "Uninstall" (the very bottom button). Say OK and allow the system to reboot. I know, it seems counterintuitive and maybe a little scary to just uninstall this thing... how will the system know what to use when you restart it? This you have to take on faith. Why doesn't Microsoft explain here what XP is about to do? I dunno. But hold your breath and do it anyway. When you reboot your pc, Windows XP will notice you don't have this driver installed, and it will select one for you automatically. It won't even ask and it won't even tell you until your desktop loads and there, you should see one of those goofy little bubbles in your system tray that says something like "XP discovered new hardware (your diskname here) and installed it for you." "What!??" you say? "I didn't install any new hard disk!" Heh. I know, but that's what XP says when it finds and re-installs the driver you removed. You should also notice that your pc suddenly works a lot faster and that Process Explorer (or even the simplistic windows taskmanager.exe) is not showing gigantic amounts of CPU kernel time being dedicated to reading hardware interrupts. Now if you go back through the above steps, Device Manager may show that "nothing" is selected in Box three, where you'd expect it to see "Ultra DMA Mode 5" or the like. This will show up when you reboot again and, barring other problems, all will be right with your world again. Thanks for this great forum and for excellent tools like Process Explorer! |
|
![]() |
|
Post Reply
|
Page <1 4567> |
|
Tweet
|
| Forum Jump | Forum Permissions ![]() You cannot post new topics in this forum You cannot reply to topics in this forum You cannot delete your posts in this forum You cannot edit your posts in this forum You cannot create polls in this forum You cannot vote in polls in this forum |