GPU-Z Video card GPU Information Utility

I am looking for an application in Linux that will give me Graphic card sensor data.
Something like TechPowerUp GPU-Z https://www.techpowerup.com/gpuz/

Could not find many alternatives:
http://alternativeto.net/software/gpu-z/?platform=linux

Is it possible to get information in Linux about graphic card as displayd in TechPowerUp GPU-Z. Perhaps through some low level utility, linux tools or parsing files. If the information is available I was thinking of making a small program for displaying them.
http://www.techpowerup.com/forums/attachments/uefi-jpg.65189/

Only thing that comes to mind is nvidia-settings program if you installed the proprietary driver. There is equivalents program that comes with the AMD proprietary driver.

hwinfo will parse various information from /proc, files, etc. and return a nice summary. You can query against specific system components like:

hwinfo --gfxcard

There are existing GUI front ends for hwinfo.

However, this won’t provide all the information like gpu-z does - and no sensor data.

For Nvidia devices, nvidia-settings will provide some sensor information, see nvidia-settings -q all for all the possible query values, there are quite a few.

There is also nvidia-smi which can provided additional information, including sensor data:

nvidia-smi -q -d temperature | grep GPU

Also for Nvidia there was a utility called http://www.linuxhardware.org/nvclock/ though development stopped a good while back the source is still available, and might be useful.

You could also explore lmsensor and it’s various graphical front ends, like http://wpitchoune.net/blog/psensor/

Command nvidia-settings -q all spewed out a lot of data. Not many attributes there are usefull or interessting.
nvidia-settings -q -d NvidiaDriverVersion
nvidia-settings -q -d OpenGLVersion

PCIEMaxLinkSpeed: 5000
PCIEMaxLinkWidth: 16
PCIECurrentLinkSpeed: 2500
PCIECurrentLinkWidth: 16
VideoRam: 1048576
TotalDedicatedGPUMemory: 1016
UsedDedicatedGPUMemory: 651
CUDACores: 768
GPUMemoryInterface: 128
GPUCoreTemp: 41
GPUCurrentCoreVoltage: 0
GPUCurrentClockFreqs: 324,324.
GPUCurrentFanSpeedRPM: 930
BusRate: 16
PCIEGen: 2
ThermalSensorReading: 41

Only attribute I could find regarding SLI was SliMosaicModeAvailable.

Othwewise nvidia-smi looks to offer some information too, but a lot of it is N/A.
Driver Version : 361.28
Attached GPUs : 1
GPU 0000:01:00.0
Product Name : GeForce GTX 650 Ti

I reckon perhaps that GPU-Z is also using some nvidia tools to get information about the graphic card and sensor data.

DjViking, the majority of information GPU-Z shows is just static database values for your graphics card :). This can be found online, even on nVidia’s website. The dynamic stuff you might be interested in (temperatures, current clock rates, memory usage, etc.) is all provided by the nvidia-settings GUI panel.

However, if you’re not using the proprietary driver, that’s a completely different story. Outputs from nouveau monitoring tools are very technical and inaccessible to average users.

Does GPU-Z has its own database online or is it querying Nvidia websites for the extra information?

Because of the lack of applications to display GPU information and sensor data (at least for KDE and what I have found) I have started my own project.
I will start with a small GUI to show some of the same information that GPU-Z provides. Using nvidia-settings and nvidia-smi

The GPU-Z information about Bus Interface: PCI-E 3.0 x16@x16 1.1 it that taken from the OS?
Using nvidia-smi I could see some information about PCI Bus, but it is not human readable
PCI
Bus : 0x01
Device : 0x00

I could also use the suggested command “hwinfo --gfxcard” to get some basic information about the graphic card and nvidia-settings for the rest.

Why does so many attributes in nvidia-smi show N/A?

  1. There would be no point in GPU-Z having its own database when all of the information can be queried directly from nVidia. All it does is identification of your GPU and displaying commonly accessible data for that GPU.

  2. Why would you need more applications for displaying GPU information and sensor data when you already have the nvidia-settings GUI and the AMD/ATI Catalyst panel?

  3. The PCI-E bus interface information from nvidia-settings is presented in a human-readable form.

  4. The N/A in nvidia-smi might mean that the sensors don’t have a particular functionality, the drivers don’t support the particular functionality or that the GPU itself doesn’t have it.

  1. I guess one could parse information from GeForce GTX 650 Ti | Specifications | GeForce
  2. Mostly for fun. Not having to memorize all the attribute names for easy access (nvidia-settings -t -q [gpu:0]/GPUCoreTemp)
  3. So really Bus 0x01 is human readable? What does 0x01 stand for otherwise equals to 1

Hi
On my system with a 8800GT, I use conky and the following commands;


Graphics Clock - nvidia-settings -t -q [gpu:0]/GPUCurrentClockFreqs}
GPU Temperature - nvidia-settings -t -q [thermalsensor:0]/ThermalSensorReading}.0
GPU Board Temperature - nvidia-settings -t -q [thermalsensor:1]/ThermalSensorReading}.0
GPU Fan - nvidia-settings -t -q GPUCurrentFanSpeed}.0

I also have a script to control the fan speed via nvidia-settings using for example;


nvidia-settings -a [gpu:0]/GPUFanControlState=1 -a [fan:0]/GPUCurrentFanSpeed=30