I have stopped using multiprocessors and IO APIC in VirtualBox.
I especial noted a sudden slow down of e.g. screen response when using some programs, which disappeared after disabling IO APIC.
Disk performance and multiple cores
-
- Posts: 20
- Joined: 3. Nov 2009, 10:00
- Primary OS: MS Windows 2003
- VBox Version: PUEL
- Guest OSses: Windows XP
Re: Disk performance and multiple cores
Thanks for your suggestion. I think it's worth a try, but as far as I understand it is recommended to enable this option for modern operating systems?
-
- Site Moderator
- Posts: 39156
- Joined: 4. Sep 2008, 17:09
- Primary OS: MS Windows 10
- VBox Version: PUEL
- Guest OSses: Mostly XP
Re: Disk performance and multiple cores
XP works great as a guest without the IO APIC option set. In fact, subjectively it works better.
Be careful however. If the XP guest was installed with "IO APIC" on, and you simply turn it off, then the XP guest will hang on a black screen during the boot. This is because the XP guest is using the wrong hardware abstraction layer. The HALu utility can be used to change to a simpler HAL (ACPI PC is fine if you don't need multiple cores in the guest), then the IO APIC option can safely be switched off.
Be careful however. If the XP guest was installed with "IO APIC" on, and you simply turn it off, then the XP guest will hang on a black screen during the boot. This is because the XP guest is using the wrong hardware abstraction layer. The HALu utility can be used to change to a simpler HAL (ACPI PC is fine if you don't need multiple cores in the guest), then the IO APIC option can safely be switched off.
-
- Posts: 20
- Joined: 3. Nov 2009, 10:00
- Primary OS: MS Windows 2003
- VBox Version: PUEL
- Guest OSses: Windows XP
Re: Disk performance and multiple cores
I just tried to boot my VM without I/O APIC and two cores, it worked. The performance tests for disk I/O was much better...
I'm wondering why I did not have to change the HAL...
I'm wondering why I did not have to change the HAL...
-
- Posts: 583
- Joined: 20. Oct 2009, 01:22
- Primary OS: MS Windows 10
- VBox Version: PUEL
- Guest OSses: Windows, LinuxMint
Re: Disk performance and multiple cores
If you disable IO APIC you will only get one processor in XP - what-so-ever!
But as said, when XP runs much better with one processor, why have more!
But as said, when XP runs much better with one processor, why have more!
-
- Posts: 20
- Joined: 3. Nov 2009, 10:00
- Primary OS: MS Windows 2003
- VBox Version: PUEL
- Guest OSses: Windows XP
Re: Disk performance and multiple cores
You're right, the Virtual Box GUI reenables the option I/O APIC when changing to two cores even if you disabled it before. I missed that!
Now I also get the black screen when booting... So I change back to one core for the moment.
It would be nice to have more than one core if it would be faster
Now I also get the black screen when booting... So I change back to one core for the moment.
It would be nice to have more than one core if it would be faster
-
- Posts: 27
- Joined: 21. Mar 2009, 00:17
- Primary OS: Debian other
- VBox Version: PUEL
- Guest OSses: Win XP
Re: Disk performance and multiple cores
Would you care to comment on the procedure used for measuring disk speed vs number of cores in this thread?mpack wrote:@stkris - the OP talked specifically about disk performance, you are only confusing matters here. You may or may not be correct that using multiple cores slows things down, but that doesn't change the fact that using tools such Passmark is a waste of time inside a VM, and results from such a tool will not be accepted as useful evidence.
http://forums.virtualbox.org/viewtopic.php?f=7&t=26611
-
- Site Moderator
- Posts: 39156
- Joined: 4. Sep 2008, 17:09
- Primary OS: MS Windows 10
- VBox Version: PUEL
- Guest OSses: Mostly XP
Re: Disk performance and multiple cores
Admittedly after just a quick read, I see no procedure mentioned in that thread. I see only reported observations (and those not directly about disk speed). Nor do I have much interest in how fast Visual Studio runs inside a multi-core VM, the only issue I have is with people reporting the results of benchmarking apps which were run inside a VM (or equivalently: the applications own reported performance). If some truly independent method of measuring time is used then fine.stkris wrote:Would you care to comment on the procedure used for measuring disk speed vs number of cores in this thread?
-
- Posts: 2
- Joined: 27. Aug 2009, 05:13
- Primary OS: Ubuntu 8.04
- VBox Version: OSE Debian
- Guest OSses: Ubuntu
Re: Disk performance and multiple cores
Sorry about thread necromancy but I just wanted to find out if this issue has had any light shed on it since 2010.
I have had difficulty with Disk I/O performance with a Windows 2008 SR2 guest on a linux server.
Here are the details of guest and host...
I've been testing Windows 2008 SR2 guest with the following:
* VirtualBox 4.1.2
* 200 gig VDI fixed size disk image
* 16gig of RAM
* 12 x CPU cores
The host machine is a Debian server with 2 x Xeon E5620 CPU, SSD disks.
* Debian lenny
* 2 x Xeon E5620 (effectively 16 cores)
* 24gig ram
* 4 x 120gig SSD's running in a RAID5 configuration.
My problem is that it seems the more Cores I enable on the Guest the slower the disk performance is.
Here is the disk performance on the guest running with 10 cores enabled using a dumb little disk tool called CrystalDiskMark.
Here is it running with 6 cores enabled.
Here it is with 2 cores enabled.
And for reference here is a small Windows XP guest machine on the same Host server with a single core.
Here is the Virutalbox Windows 2008 SR2 Guest config with 12 CPU Core's enabled.
I have had difficulty with Disk I/O performance with a Windows 2008 SR2 guest on a linux server.
Here are the details of guest and host...
I've been testing Windows 2008 SR2 guest with the following:
* VirtualBox 4.1.2
* 200 gig VDI fixed size disk image
* 16gig of RAM
* 12 x CPU cores
The host machine is a Debian server with 2 x Xeon E5620 CPU, SSD disks.
* Debian lenny
* 2 x Xeon E5620 (effectively 16 cores)
* 24gig ram
* 4 x 120gig SSD's running in a RAID5 configuration.
My problem is that it seems the more Cores I enable on the Guest the slower the disk performance is.
Here is the disk performance on the guest running with 10 cores enabled using a dumb little disk tool called CrystalDiskMark.
Here is it running with 6 cores enabled.
Here it is with 2 cores enabled.
And for reference here is a small Windows XP guest machine on the same Host server with a single core.
Here is the Virutalbox Windows 2008 SR2 Guest config with 12 CPU Core's enabled.
Code: Select all
Name: cssp
Guest OS: Windows 2008 (64 bit)
UUID: 2c4d6076-0600-409f-b0d7-107db135b4e0
Config file: /home/sysadmin/.VirtualBox/Machines/cssp/cssp.xml
Snapshot folder: /home/sysadmin/.VirtualBox/Machines/cssp/Snapshots
Log folder: /home/sysadmin/.VirtualBox/Machines/cssp/Logs
Hardware UUID: 2c4d6076-0600-409f-b0d7-107db135b4e0
Memory size: 16384MB
Page Fusion: off
VRAM size: 16MB
CPU exec cap: 100%
HPET: off
Chipset: piix3
Firmware: BIOS
Number of CPUs: 12
Synthetic Cpu: off
CPUID overrides: None
Boot menu mode: message and menu
Boot Device (1): Floppy
Boot Device (2): DVD
Boot Device (3): HardDisk
Boot Device (4): Not Assigned
ACPI: on
IOAPIC: on
PAE: off
Time offset: 0 ms
RTC: local time
Hardw. virt.ext: on
Hardw. virt.ext exclusive: on
Nested Paging: off
Large Pages: off
VT-x VPID: on
State: powered off (since 2011-09-13T10:10:57.655000000)
Monitor count: 1
3D Acceleration: off
2D Video Acceleration: off
Teleporter Enabled: off
Teleporter Port: 0
Teleporter Address:
Teleporter Password:
Storage Controller Name (0): IDE Controller
Storage Controller Type (0): PIIX4
Storage Controller Instance Number (0): 0
Storage Controller Max Port Count (0): 2
Storage Controller Port Count (0): 2
Storage Controller Bootable (0): on
Storage Controller Name (1): Floppy Controller
Storage Controller Type (1): I82078
Storage Controller Instance Number (1): 0
Storage Controller Max Port Count (1): 1
Storage Controller Port Count (1): 1
Storage Controller Bootable (1): on
Storage Controller Name (2): SATA Controller
Storage Controller Type (2): IntelAhci
Storage Controller Instance Number (2): 0
Storage Controller Max Port Count (2): 30
Storage Controller Port Count (2): 3
Storage Controller Bootable (2): on
IDE Controller (1, 0): /usr/share/virtualbox/VBoxGuestAdditions.iso (UUID: 10efa78a-1f35-4431-a8bf-12c8130c0bd0)
SATA Controller (0, 0): /home/sysadmin/.VirtualBox/HardDisks/cssp.vdi (UUID: 3d445880-5a1d-413d-a676-7593c1c34425)
SATA Controller (2, 0): /home/sysadmin/.VirtualBox/HardDisks-nas/cssp-backup.vdi (UUID: b02a4236-0b56-4694-911f-ac2aad129332)
NIC 1: MAC: 08002760E77C, Attachment: Bridged Interface 'eth0', Cable connected: on, Trace: off (file: none), Type: 82540EM, Reported speed: 0 Mbps, Boot priority: 0, Promisc Policy: deny
NIC 2: disabled
NIC 3: disabled
NIC 4: disabled
NIC 5: disabled
NIC 6: disabled
NIC 7: disabled
NIC 8: disabled
Pointing Device: PS/2 Mouse
Keyboard Device: PS/2 Keyboard
UART 1: I/O base: 0x03f8, IRQ: 4, attached to device '/dev/ttyS0'
UART 2: disabled
Audio: enabled (Driver: Null, Controller: HDA)
Clipboard Mode: Bidirectional
VRDE: disabled
USB: disabled
USB Device Filters:
Index: 0
Active: yes
Name: Agere Systems Agere USB2.0 V.92 SoftModem [0200]
VendorId: 047e
ProductId: 2892
Revision: 0200
Manufacturer: Agere Systems
Product: Agere USB2.0 V.92 SoftModem
Remote: 0
Serial Number:
Available remote USB devices:
<none>
Currently Attached USB Devices:
<none>
Shared folders: <none>
VRDE Connection: not active
Clients so far: 0
Guest:
Configured memory balloon size: 0 MB
OS type: Windows2008_64
Additions run level: 0
Guest Facilities:
No active facilities.