NVIDIA graphics card not recognised

[jetojedno]

I have two NVIDIA gfx cards in my machine, and the second card isn’t being recognised / having the driver loaded.

here is the relevant section of lspci -v (compared using diff --side-by-side) with just the “invisible” card" installed on the left, and both cards on the right.

**03:00.0** VGA compatible controller: NVIDIA Corporation Device  | **01:00.0** VGA compatible controller: NVIDIA Corporation GF108GL
        Subsystem: NVIDIA Corporation **Device 1c82**             |         Subsystem: NVIDIA Corporation **Device 0835**
        **Physical Slot: 4 **                                     |         **Physical Slot: 2**
        Flags: bus master, fast devsel, latency 0, IRQ 16     |         Flags: bus master, fast devsel, latency 0, IRQ 31
        Memory at f9000000 (32-bit, non-prefetchable) [size=1] |         Memory at d0000000 (64-bit, prefetchable) [size=128]
        Memory at d0000000 (64-bit, prefetchable) [size=32]  |         Memory at d8000000 (64-bit, prefetchable) [size=32]
        I/O ports at e000 [size=128]                                    I/O ports at e000 [size=128]
        Expansion ROM at fa000000 [disabled] [size=512]      |         [virtual] Expansion ROM at fa000000 [disabled] [size=] Power Management version 3                   Capabilities: [60] Power Management version 3
        Capabilities: [68] MSI: Enable- Count=1/1 Maskable- 6 |         Capabilities: [68] MSI: Enable+ Count=1/1 Maskable- 6
        Capabilities: [78] Express Legacy Endpoint, MSI 00    |         Capabilities: [78] Express Endpoint, MSI 00
                                                              >         Capabilities: [b4] Vendor Specific Information: Len=1
        Capabilities: [100] Virtual Channel                             Capabilities: [100] Virtual Channel
        Capabilities: [250] Latency Tolerance Reporting       <
        Capabilities: [128] Power Budgeting <?>                         Capabilities: [128] Power Budgeting <?>
        Capabilities: [420] Advanced Error Reporting          <
        Capabilities: [600] Vendor Specific Information: ID=0           Capabilities: [600] Vendor Specific Information: ID=0
        Capabilities: [900] #19                               |         Kernel driver in use: nvidia
        Kernel modules: nouveau                               |         Kernel modules: nouveau, nvidia_drm, nvidia


03:00.1 Audio device: NVIDIA Corporation Device 0fb9 (rev a1) | 01:00.1 Audio device: NVIDIA Corporation GF108 High Definitio
        Subsystem: NVIDIA Corporation Device 1c82             |         Subsystem: NVIDIA Corporation Device 0835
        Physical Slot: 4                                      |         Physical Slot: 2
        Flags: bus master, fast devsel, latency 0, IRQ 17               Flags: bus master, fast devsel, latency 0, IRQ 17
        Memory at fa080000 (32-bit, non-prefetchable) [size=1] Power Management version 3                   Capabilities: [60] Power Management version 3
        Capabilities: [68] MSI: Enable- Count=1/1 Maskable- 6           Capabilities: [68] MSI: Enable- Count=1/1 Maskable- 6
        Capabilities: [78] Express Endpoint, MSI 00                     Capabilities: [78] Express Endpoint, MSI 00
        Capabilities: [100] Advanced Error Reporting          <
        Kernel driver in use: snd_hda_intel                             Kernel driver in use: snd_hda_intel
        Kernel modules: snd_hda_intel                Kernel modules: snd_hda_intel

Can you advise:

1. what additional evidence / logs etc would help locate the failure, and/or
2. how I can “force” recognition of the invisible card & load the driver?

Thanks in advance,
David[/size][/size][/size][/size][/size][/size][/size][/size][/size]

[knurpht]

Well the left output does “see” the card, even has the nouveau driver loaded. I haven’t had any NVIDIA graphics based systems at hand, but IIRC there’s an option in the NVIDIA installer to handle multiple cards, but I’m not sure.

[jetojedno]

@Knurpht

there’s an option in the NVIDIA installer to handle multiple cards

Unfortunately:
http://www.tomshardware.co.uk/answers/id-3229891/nvidia-drivers-unstable-multiple-cards.html

I think your problem may stem from trying to mix and match cards from different series. If you tried three cards from the same generation you might have better luck.

[jetojedno]

@Knurpht

Well the left output does “see” the card, even has the nouveau driver loaded. I haven’t had any NVIDIA graphics based systems at hand, but IIRC there’s an option in the NVIDIA installer to handle multiple cards, but I’m not sure.

I forgot to say thank you … your comment pointed me in the right direction and some quick searching found an answer - the two cards won’t mix.

I’ve sorted the situation by swapping one of the NVIDIA cards for an ATI one, and now both cards are seen and respond.

Sometimes it’s useful having a bit of bits stripped from old machines …

[bryga]

It looks like nouveau and nvidia proprietary drivers were loaded. They usually do not mix well, especially when you have 2 or more cards. Having different cards is not a problem as long as you do not want to use them in SLI mode.
You can try blacklist (add) nouveau driver in /etc/modprobe.d/50-blacklist.conf and run mkinitrd. And reboot. Then only nvidia proprietary driver will get loaded.

To check if nvidia and/or nouveau drivers loaded run (probably as root):
lsmod | egrep “nvdia|nouveau”

@jetojedno

I forgot to say thank you … your comment pointed me in the right direction and some quick searching found an answer - the two cards won’t mix.

I’ve sorted the situation by swapping one of the NVIDIA cards for an ATI one, and now both cards are seen and respond.

Sometimes it’s useful having a bit of bits stripped from old machines …

[jetojedno]

@bryga

It looks like nouveau and nvidia proprietary drivers were loaded.

Sorry, I should have been clearer … I did quite a bit of messing round trying to get both cards seen by the nvidia driver, including reverting to the nouveau driver. The lspci with the single card was from when I had the nouveau driver loaded & had taken out one of the cards to see if it was a hardware problem (it wasn’t). I then switched to the nvidia driver (needed for CUDA) and then added back the second card (first card disappeared), which is when I saved the next lspci.

Now, the two cards are present and working …

$ /sbin/lspci -nn
01:00.0 VGA compatible controller [0300]: Advanced Micro Devices, Inc. [AMD/ATI] RV370 [Radeon X300/X550/X1050 Series] [1002:5b63]
01:00.1 Display controller [0380]: Advanced Micro Devices, Inc. [AMD/ATI] RV370 [Radeon X300/X550/X1050 Series] (Secondary) [1002:5b73]
03:00.0 VGA compatible controller [0300]: NVIDIA Corporation GP107 [GeForce GTX 1050 Ti] [10de:1c82] (rev a1)
03:00.1 Audio device [0403]: NVIDIA Corporation GP107GL High Definition Audio Controller [10de:0fb9] (rev a1)

$ lsmod | egrep -i 'nvidia|nouveau|gfx|vga'
nvidia_drm             49152  0 
nvidia_modeset        860160  2 nvidia_drm
drm_kms_helper        155648  1 nvidia_drm
drm                   393216  4 ttm,drm_kms_helper,nvidia_drm
nvidia_uvm            704512  0 
nvidia              13160448  10 nvidia_modeset,nvidia_uvm

$ deviceQuerydeviceQuery Starting...


 CUDA Device Query (Runtime API) version (CUDART static linking)


Detected 1 CUDA Capable device(s)


Device 0: "GeForce GTX 1050 Ti"
  CUDA Driver Version / Runtime Version          9.0 / 9.0
  CUDA Capability Major/Minor version number:    6.1
  Total amount of global memory:                 4039 MBytes (4235001856 bytes)
  ( 6) Multiprocessors, (128) CUDA Cores/MP:     768 CUDA Cores
  GPU Max Clock rate:                            1392 MHz (1.39 GHz)
  Memory Clock rate:                             3504 Mhz
  Memory Bus Width:                              128-bit
  L2 Cache Size:                                 1048576 bytes
  Maximum Texture Dimension Size (x,y,z)         1D=(131072), 2D=(131072, 65536), 3D=(16384, 16384, 16384)
  Maximum Layered 1D Texture Size, (num) layers  1D=(32768), 2048 layers
  Maximum Layered 2D Texture Size, (num) layers  2D=(32768, 32768), 2048 layers
  Total amount of constant memory:               65536 bytes
  Total amount of shared memory per block:       49152 bytes
  Total number of registers available per block: 65536
  Warp size:                                     32
  Maximum number of threads per multiprocessor:  2048
  Maximum number of threads per block:           1024
  Max dimension size of a thread block (x,y,z): (1024, 1024, 64)
  Max dimension size of a grid size    (x,y,z): (2147483647, 65535, 65535)
  Maximum memory pitch:                          2147483647 bytes
  Texture alignment:                             512 bytes
  Concurrent copy and kernel execution:          Yes with 2 copy engine(s)
  Run time limit on kernels:                     No
  Integrated GPU sharing Host Memory:            No
  Support host page-locked memory mapping:       Yes
  Alignment requirement for Surfaces:            Yes
  Device has ECC support:                        Disabled
  Device supports Unified Addressing (UVA):      Yes
  Supports Cooperative Kernel Launch:            Yes
  Supports MultiDevice Co-op Kernel Launch:      Yes
  Device PCI Domain ID / Bus ID / location ID:   0 / 3 / 0
  Compute Mode:
     < Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) >


deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 9.0, CUDA Runtime Version = 9.0, NumDevs = 1
Result = PASS

… so, I’m now happy enough … until I break something …

[jetojedno]

I should add that I’ve just started a Deep Learning course, and this is to save me a small fortune in renting cloud VMs + GPUs for DNN training …