You are not logged in.

#1 2017-12-13 22:09:23

Walton
Member
Registered: 2014-01-09
Posts: 20

[Minor bug] lspci and lshw shows wrong graphics card (GTX 1060)

I don't know if this will cause performance issues, or this is just a bug, but lspci and lshw are showing the the wrong graphics card. I have a Dell Inspiron 7000 gaming edition.

[ steven ] [~] > lspci | grep -i nvidia                                                                                                                                                   []
01:00.0 VGA compatible controller: NVIDIA Corporation GP106M [GeForce GTX 1060 Mobile 3GB] (rev a1)
[ steven ] [~] > lshw
...
*-display
                description: VGA compatible controller
                product: GP106M [GeForce GTX 1060 Mobile 3GB]
                vendor: NVIDIA Corporation
                physical id: 0
                bus info: pci@0000:01:00.0
                version: a1
                width: 64 bits
                clock: 33MHz
                capabilities: vga_controller bus_master cap_list rom
                configuration: driver=nvidia latency=0
                resources: irq:16 memory:dc000000-dcffffff memory:b0000000-bfffffff memory:c0000000-c1ffffff ioport:e000(size=128) memory:dd000000-dd07ffff
...

Yet, when I run the Mandlebrot example code from the cuda samples I get

[ steven ] [~/Programming/testingStuff/CudaSamples/2_Graphics/Mandelbrot] > ./Mandelbrot                                                                                                  []
[CUDA Mandelbrot/Julia Set] - Starting...
GPU Device 0: "GeForce GTX 1060 with Max-Q Design" with compute capability 6.1

Data initialization done.
Initializing GLUT...
OpenGL window created.
Creating GL texture...
Texture created.
Creating PBO...
CUDA error at Mandelbrot.cpp:972 code=46(cudaErrorDevicesUnavailable) "cudaGraphicsGLRegisterBuffer(&cuda_pbo_resource, gl_PBO, cudaGraphicsMapFlagsWriteDiscard)" 

I know that the correct card is the 1060 Max-Q, with 6 GB. My Windows side confirms this as well, matching the listing from costco (where I bought it). What is more interesting, is that I'm not even aware of a mobile version of this card existing. SMI also shows the correct hardware (similar results with the q flag)

[ steven ] [~] > nvidia-smi                                                                                                                                                               []
Wed Dec 13 16:01:02 2017       
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 387.34                 Driver Version: 387.34                    |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce GTX 106...  Off  | 00000000:01:00.0 Off |                  N/A |
| N/A   52C    P0    21W /  N/A |      0MiB /  6072MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID   Type   Process name                             Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+

I am using the proprietary nvidia drivers.

[ steven ] [~] > pacman -Qs nvidia                                                                                                                                                        []
local/bumblebee 3.2.1-15
    NVIDIA Optimus support for Linux through VirtualGL
local/cuda 9.0.176-4
    NVIDIA's GPU programming toolkit
local/libvdpau 1.1.1+3+ga21bf7a-1
    Nvidia VDPAU library
local/nvidia 387.34-11
    NVIDIA drivers for linux
local/nvidia-utils 387.34-1
    NVIDIA drivers utilities
local/opencl-nvidia 387.34-1
    OpenCL implemention for NVIDIA

Is this a known bug? Is it going to affect anything? Or is this just a harmless thing?

Last edited by Walton (2017-12-13 22:59:45)

Offline

#2 2017-12-13 22:52:07

progandy
Member
Registered: 2012-05-17
Posts: 2,786

Re: [Minor bug] lspci and lshw shows wrong graphics card (GTX 1060)

That is really harmless, just a minor documentation error. The card reports an identifier that is translated into a human readable format with a database. This identifier might be used for slightly different implementations of the same graphics chip. New versions of the database have been updated and the "3GB" has been removed[1]. The next update of [core]/hwids will have it, but that might still take a while since arch pulls its copy from gentoo where it can take up to a year to get a new database [2].

1: https://github.com/pciutils/pciids/comm … 93c8c77597
2: https://github.com/gentoo/hwids/commits/master

Offline

#3 2017-12-13 23:05:43

Walton
Member
Registered: 2014-01-09
Posts: 20

Re: [Minor bug] lspci and lshw shows wrong graphics card (GTX 1060)

Well that is good to know. I am marking the thread as a minor bug.

This may just be my unfamiliarity with this type of stuff, but in the first documentation is still looks like those are reporting a mobile graphics card (I notice this in more than the 1060), which don't exist in the GTX series. And shouldn't there be a configuration for a 6GB, like there is in the non-Max-Q version?

Offline

#4 2017-12-13 23:21:45

progandy
Member
Registered: 2012-05-17
Posts: 2,786

Re: [Minor bug] lspci and lshw shows wrong graphics card (GTX 1060)

The list is maintained by volunteers here:
https://pci-ids.ucw.cz/

Offline

Board footer

Powered by FluxBB