[SOLVED] Issue installing a external GPU device

aihaike · 07-04-2011, 12:35 AM

Oups!
Here comes the picture of the system I mentioned previously.

EDDY1 · 07-04-2011, 12:36 AM

Quote:

I'm sorry, but I don't see why lsusb would be helpful here.
It seems there is a confusion so I've attached a picture of what looks like my system where
PE4H is connected to the card and HP4A to the express PCI slot.
I do not use any USB connexion, that's what I put the output of lspci.

My apologies you are correct it was "lspci", which you provided & also isn't recognizing PH4A

aihaike · 07-04-2011, 12:46 AM

You're right, lspci has to show it.
So, I have to focus on that first.
Thank you,

Éric.

EDDY1 · 07-04-2011, 12:50 AM

I am actively looking for firmware though

aihaike · 07-04-2011, 02:52 AM

Here are some more informations about my system:

I've just discovered biosdecode which gives

Code:

# biosdecode 2.10
SYSID present.
	Revision: 0
	Structure Table Address: 0x000F0411
	Number Of Structures: 1
SMBIOS 2.5 present.
	Structure Table Length: 3013 bytes
	Structure Table Address: 0x000F0450
	Number Of Structures: 82
	Maximum Structure Size: 253 bytes
PNP BIOS 1.0 present.
	Event Notification: Not Supported
	Real Mode 16-bit Code Address: F000:E2F1
	Real Mode 16-bit Data Address: 0040:0000
	16-bit Protected Mode Code Address: 0x000FE2F4
	16-bit Protected Mode Data Address: 0x00000040
ACPI 2.0 present.
	OEM Identifier: DELL  
	RSD Table 32-bit Address: 0x000FC75B
	XSD Table 64-bit Address: 0x00000000000FC7CB
PCI Interrupt Routing 1.0 present.
	Router ID: 00:1f.0
	Exclusive IRQs: None
	Compatible Router: 8086:3a14
	Slot Entry 1: ID 00:01, on-board
	Slot Entry 2: ID 00:02, on-board
	Slot Entry 3: ID 00:03, on-board
	Slot Entry 4: ID 00:16, on-board
	Slot Entry 5: ID 00:17, on-board
	Slot Entry 6: ID 00:1a, on-board
	Slot Entry 7: ID 00:1b, on-board
	Slot Entry 8: ID 00:1c, on-board
	Slot Entry 9: ID 00:1d, on-board
	Slot Entry 10: ID 00:1f, on-board
	Slot Entry 11: ID 01:00, slot number 10
	Slot Entry 12: ID 00:19, on-board
	Slot Entry 13: ID 03:00, slot number 4
	Slot Entry 14: ID 02:00, slot number 5
	Slot Entry 15: ID 04:01, slot number 1
	Slot Entry 16: ID 04:00, slot number 2
	Slot Entry 17: ID 04:02, slot number 3
BIOS32 Service Directory present.
	Revision: 0
	Calling Interface Address: 0x000FFEA0

and dmidecode for System Slots (t -9 option)

Code:

# dmidecode 2.10
SMBIOS 2.5 present.

Handle 0x090A, DMI type 9, 13 bytes
System Slot Information
	Designation: SLOT1
	Type: x1 Proprietary
	Current Usage: Available
	Length: Long
	Characteristics:
		PME signal is supported

Handle 0x0902, DMI type 9, 13 bytes
System Slot Information
	Designation: SLOT2  
	Type: 32-bit PCI
	Current Usage: Available
	Length: Long
	ID: 2
	Characteristics:
		5.0 V is provided
		3.3 V is provided
		PME signal is supported

Handle 0x0903, DMI type 9, 13 bytes
System Slot Information
	Designation: SLOT3  
	Type: 32-bit PCI
	Current Usage: Available
	Length: Long
	ID: 3
	Characteristics:
		5.0 V is provided
		3.3 V is provided
		PME signal is supported

Handle 0x0904, DMI type 9, 13 bytes
System Slot Information
	Designation: SLOT4  
	Type: x1 PCI Express
	Current Usage: Available
	Length: Long
	ID: 4
	Characteristics:
		3.3 V is provided
		PME signal is supported

aihaike · 07-04-2011, 08:31 PM

From what I read, it seems difficult to install the GPU without using it as the display device (with my intel onboard graphics as the main display) ...

aihaike · 07-04-2011, 09:15 PM

ok, so it seems I managed it to work.
The point is I don't quit understand how I did it.
I rebooted the machine with the minitor connectede to the card, just to see.
Well, I saw nothing actually, so I rebbooted with the monitor connected on my onboard card but forgot to unplug the HDMI to VGA adptor from the card.
Now I'm wondering whether the card is detecded because I somehaow "foced" it to be detecded by connecteing it to the minitor or because there is still the adaptor.
I don't dare take it out ...
Here is what "lspci -vv" reports :

Code:

01:00.0 VGA compatible controller: nVidia Corporation Device 1080 (rev a1) (prog-if 00 [VGA controller])
	Subsystem: Device 196e:086a
	Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
	Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
	Latency: 0, Cache Line Size: 64 bytes
	Interrupt: pin A routed to IRQ 16
	Region 0: Memory at fc000000 (32-bit, non-prefetchable) [size=16M]
	Region 1: Memory at d0000000 (64-bit, prefetchable) [size=128M]
	Region 3: Memory at d8000000 (64-bit, prefetchable) [size=32M]
	Region 5: I/O ports at dc80 [disabled] [size=128]
	Expansion ROM at fd900000 [disabled] [size=512K]
	Capabilities: <access denied>
	Kernel driver in use: nvidia
	Kernel modules: nvidia, nouveau, nvidiafb

Actually the name of the card (GTX 280) is not shown

And "egrep -i 'nvidia|NVIDIA'" gives

Code:

nvidia: module license 'NVIDIA' taints kernel.
nvidia 0000:01:00.0: enabling device (0002 -> 0003)
nvidia 0000:01:00.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16
nvidia 0000:01:00.0: setting latency timer to 64
NVRM: loading NVIDIA UNIX x86_64 Kernel Module  260.19.29  Wed Dec  8 12:08:56 PST 2010

Just need to run some test now ...

cascade9 · 07-05-2011, 01:12 AM

Quote:

Originally Posted by aihaike

Oups!
Here comes the picture of the system I mentioned previously.

Ummm...that looks like you have a PCIe x16 slot. Why not just install the GTX280/GTX580 into the PCIe slot on the board? (I would guess that GTX580 is a typo, it looks more like a GTX280 to me)

EDDY1 · 07-05-2011, 01:42 AM

I myself hope you figure out what you did to get it recognized so others can benefit from it as from the threads I've seen there are only questions with no solutions. Like I said I'm not the 1 here that is experienced with your operation, but there are others here, that can help you figure it out(what you did to get it working) & create a formula for the next person using your (I guess) bash history.
congrats on getting this far.

aihaike · 07-05-2011, 01:54 AM

Well, it's not done yet.
I had to upgrade the module, so I try to do a "rmmod nvidia", compile the new one, and atempt to do a "modprob nvidia". It did not work (obviously some may say).
So, I restart the machine but either the machine boot and the card is not detected or it is and I get "Attention: Unsupported Video Configuration Detected".
The fact that the adapter is plug or not does not change anything.
So weird!!
And I was able to run CUDA program!!
I hope I did not damage anything with rmmod ....

EDDY1 · 07-05-2011, 02:02 AM

You will get it.

congrats on getting this far

aihaike · 07-07-2011, 10:03 PM

Hey,

I noticed that when I power on the PSU of the device at the lilo prompt and wait a little (it's important but don't know why), the card appears in lspci.
So here it is:

Code:

01:00.0 VGA compatible controller: nVidia Corporation Device 1080 (rev a1) (prog-if 00 [VGA controller])
        Subsystem: Device 196e:086a
        Control: I/O- Mem- BusMaster- SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Interrupt: pin A routed to IRQ 16
        Region 1: Memory at f0000000 (64-bit, prefetchable) [disabled] [size=128M]
        Region 3: Memory at f8000000 (64-bit, prefetchable) [disabled] [size=32M]
        Region 5: I/O ports at 1000 [disabled] [size=128]
        [virtual] Expansion ROM at fa000000 [disabled] [size=512K]
        Capabilities: <access denied>
        Kernel modules: nvidia, nouveau, nvidiafb

The problem is I can not load the module:

Code:

modprobe nvidia
FATAL: Error inserting nvidia (/lib/modules/2.6.33.4/kernel/drivers/video/nvidia.ko): No such device

and dmesg shows

Code:

NVRM: This PCI I/O region assigned to your NVIDIA device is invalid:
NVRM: BAR0 is 0M @ 0x0 (PCI:0000:01:00.0)
NVRM: The system BIOS may have misconfigured your GPU.
nvidia: probe of 0000:01:00.0 failed with error -1
NVRM: The NVIDIA probe routine failed for 1 device(s).
NVRM: None of the NVIDIA graphics adapters were initialized!

So, I wondere whether there is a way to reassigned PCI region for this device after boot.

by t he way, I also got that

Code:

vgaarb: device changed decodes: PCI:0000:00:02.0,olddecodes=io+mem,decodes=none:owns=io+mem
vgaarb: transferring owner from PCI:0000:00:02.0 to PCI:0000:01:00.0

and do not know what does it mean.

So, there is a little progress but ....

EDDY1 · 07-07-2011, 11:25 PM

Maybe there's a way & this is coming from an actual newbie, that you can figure out when it's getting recognized(as far as time) & get thmodule loaded after recognition.

Another thing, are you trying to load a gpu that has the same driver as your onboard?
I'm just guessing here, so bare with me.
If the 2 gpu's were the same driver, you think that woud work.
Hope it's not a dumb question.

Edit
I just looked back at your posts & saw Intel controller, there's a post running here about intel & nvidia you may want to look at. I want to say it's in laptop sectio if I find it I'll post back. Yesterday or day before.

aihaike · 07-11-2011, 07:14 PM

Hey,

so, after lot of investigation and test I did not really make any progress.
The device is detected 1 boot over 5. I mean, in average I have to do like 5 boots to get my card detected.
Here is what lspci -vvv returns (after updating the pci database pci.ids)

Code:

01:00.0 3D controller: nVidia Corporation GF110 [Geforce GTX 580] (rev a1)
	Subsystem: Device 196e:086a
	Control: I/O+ Mem+ BusMaster- SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
	Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
	Interrupt: pin A routed to IRQ 16
	Region 1: Memory at f0000000 (64-bit, prefetchable) [size=128M]
	Region 3: Memory at f8000000 (64-bit, prefetchable) [size=32M]
	Region 5: I/O ports at 1000 [size=128]
	[virtual] Expansion ROM at fa000000 [disabled] [size=512K]
	Capabilities: <access denied>
	Kernel modules: nvidia, nouveau, nvidiafb

01:00.1 Audio device: nVidia Corporation GF110 High Definition Audio Controller (rev a1)
	Subsystem: Device 196e:086a
	Control: I/O- Mem- BusMaster- SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
	Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
	Interrupt: pin B routed to IRQ 0
	Region 0: Memory at fe500000 (32-bit, non-prefetchable) [disabled] [size=16K]
	Capabilities: <access denied>

The good point is I/O+, but I'm concern about the BusMaster-
The would be great if an expert could tell me more about what's going wrong from the above information.

Here is what "lspci -tv" gives

Code:

-[0000:00]-+-00.0  Intel Corporation 4 Series Chipset DRAM Controller
           +-01.0-[01]--+-00.0  nVidia Corporation GF110 [Geforce GTX 580]
           |            \-00.1  nVidia Corporation GF110 High Definition Audio Controller
           +-02.0  Intel Corporation 4 Series Chipset Integrated Graphics Controller
           +-02.1  Intel Corporation 4 Series Chipset Integrated Graphics Controller
           +-03.0  Intel Corporation 4 Series Chipset HECI Controller
           +-03.2  Intel Corporation 4 Series Chipset PT IDER Controller
           +-03.3  Intel Corporation 4 Series Chipset Serial KT Controller
           +-19.0  Intel Corporation 82567LM-3 Gigabit Network Connection
           +-1a.0  Intel Corporation 82801JD/DO (ICH10 Family) USB UHCI Controller #4
           +-1a.1  Intel Corporation 82801JD/DO (ICH10 Family) USB UHCI Controller #5
           +-1a.2  Intel Corporation 82801JD/DO (ICH10 Family) USB UHCI Controller #6
           +-1a.7  Intel Corporation 82801JD/DO (ICH10 Family) USB2 EHCI Controller #2
           +-1b.0  Intel Corporation 82801JD/DO (ICH10 Family) HD Audio Controller
           +-1c.0-[02]--
           +-1c.1-[03]--
           +-1d.0  Intel Corporation 82801JD/DO (ICH10 Family) USB UHCI Controller #1
           +-1d.1  Intel Corporation 82801JD/DO (ICH10 Family) USB UHCI Controller #2
           +-1d.2  Intel Corporation 82801JD/DO (ICH10 Family) USB UHCI Controller #3
           +-1d.7  Intel Corporation 82801JD/DO (ICH10 Family) USB2 EHCI Controller #1
           +-1e.0-[04]--
           +-1f.0  Intel Corporation 82801JDO (ICH10DO) LPC Interface Controller
           +-1f.2  Intel Corporation 82801 SATA RAID Controller
           \-1f.3  Intel Corporation 82801JD/DO (ICH10 Family) SMBus Controller

I tried to set some boot option "pci=..." but did not work
I also tried to use setpci but did not quit get how does it work. Anyway, what I tried did nor work.
So that's the point.
I've also attached some picture of the set up.
I hope that may help people who wish to try to make the same stuff.

Éric.

EDDY1 · 07-11-2011, 07:43 PM

I'm not familiar with the device, but now that you're getting it to load 1 out of five times, you need to include the module associated in startup, especially if you're going to have connected to machine permanently.
I can't say exactly but I'm guessing /etc/init???