github-starred/nvitop

Fork 0

mirror of https://github.com/XuehaiPan/nvitop.git synced 2026-05-15 14:15:55 -06:00

[GH-ISSUE #168] [Feature Request] support jetson orin nano #104

New issue

Closed

opened 2026-05-05 03:25:36 -06:00 by gitea-mirror · 3 comments

gitea-mirror commented

2026-05-05 03:25:36 -06:00

Owner

Originally created by @hebais on GitHub (Jun 2, 2025).
Original GitHub issue: https://github.com/XuehaiPan/nvitop/issues/168

Originally assigned to: @XuehaiPan on GitHub.

Required prerequisites

I have searched the Issue Tracker that this hasn't already been reported. (comment there if it has.)
I have tried the latest version of nvitop in a new isolated virtual environment.

Motivation

Seems nvitop could not well detect jetson orin nano platform.
When I run nvitop command, it shows GPU MEM N/A.
My environment is:

NVITOP 1.5.1
Driver: 540.4.0
CUDA DRIVER: 12.6

By the way, I am willing to contribute to support this platform. If there is no plan to support this platform yet, I would like to dig it up. Any debug clue would be appreciated, thanks

Solution

No response

Alternatives

No response

Additional context

No response

Originally created by @hebais on GitHub (Jun 2, 2025). Original GitHub issue: https://github.com/XuehaiPan/nvitop/issues/168 Originally assigned to: @XuehaiPan on GitHub. ### Required prerequisites - [x] I have searched the [Issue Tracker](https://github.com/XuehaiPan/nvitop/issues) that this hasn't already been reported. (comment there if it has.) - [x] I have tried the latest version of nvitop in a new isolated virtual environment. ### Motivation Seems nvitop could not well detect jetson orin nano platform. When I run nvitop command, it shows GPU MEM N/A. My environment is: 1. NVITOP 1.5.1 2. Driver: 540.4.0 3. CUDA DRIVER: 12.6 By the way, I am willing to contribute to support this platform. If there is no plan to support this platform yet, I would like to dig it up. Any debug clue would be appreciated, thanks ### Solution _No response_ ### Alternatives _No response_ ### Additional context _No response_

gitea-mirror

2026-05-05 03:25:36 -06:00

closed this issue
added the
pynvml

enhancement

upstream
labels

gitea-mirror commented

2026-05-05 03:25:38 -06:00

Author

Owner

@XuehaiPan commented on GitHub (Jun 2, 2025):

When I run nvitop command, it shows GPU MEM N/A.

Hi @hebais, is there any additional context, such as screenshots or logs?

nvitop is built on top of the NVIDIA Management Library (NVML). The supported device list of NVML can be found here: https://docs.nvidia.com/deploy/nvml-api/nvml-api-reference.html#nvml-api-reference

Supported OS platforms:

Windows 64-bit: Windows Server 2019, Windows Server 2016, Windows Server 2012 R2, Windows 10

Linux: 64-bit

Supported products:

Full Support

NVIDIA Tesla Line:

A100, A40, A30, A16, A10

H100

T4

V100

P100, P40, P4, P6

M60, M40, M6, M4

K80, K520

NVIDIA Quadro Line:

RTX 8000, RTX 6000, RTX 5000, RTX 4000, RTX 3000

GV100, GP100, P6000, P5200, P5000, P4000, P2200, P2000, P1000, P620, P600, P400

M6000 24GB, M6000, M5000, M4000, M2000

K6000, K5200, K5000, K4000, K4200, K2200, K2000, K2000D, K1200, K620, K600, K420, 410

NVIDIA GeForce Line:

none

Limited Support

NVIDIA Tesla Line: All other current and previous generation Tesla-branded parts

NVIDIA Quadro Line: All other current and previous generation Quadro-branded parts

NVIDIA GeForce Line: All current and previous generation GeForce-branded parts

@XuehaiPan commented on GitHub (Jun 2, 2025): > When I run nvitop command, it shows GPU MEM N/A. Hi @hebais, is there any additional context, such as screenshots or logs? ------ `nvitop` is built on top of the [NVIDIA Management Library (NVML)](https://developer.nvidia.com/management-library-nvml). The supported device list of NVML can be found here: https://docs.nvidia.com/deploy/nvml-api/nvml-api-reference.html#nvml-api-reference > Supported OS platforms: > > - Windows 64-bit: Windows Server 2019, Windows Server 2016, Windows Server 2012 R2, Windows 10 > - Linux: 64-bit > > Supported products: > > - Full Support > - NVIDIA Tesla Line: > - A100, A40, A30, A16, A10 > - H100 > - T4 > - V100 > - P100, P40, P4, P6 > - M60, M40, M6, M4 > - K80, K520 > - NVIDIA Quadro Line: > - RTX 8000, RTX 6000, RTX 5000, RTX 4000, RTX 3000 > - GV100, GP100, P6000, P5200, P5000, P4000, P2200, P2000, P1000, P620, P600, P400 > - M6000 24GB, M6000, M5000, M4000, M2000 > - K6000, K5200, K5000, K4000, K4200, K2200, K2000, K2000D, K1200, K620, K600, K420, 410 > - NVIDIA GeForce Line: > - none > - Limited Support > - NVIDIA Tesla Line: All other current and previous generation Tesla-branded parts > - NVIDIA Quadro Line: All other current and previous generation Quadro-branded parts > - NVIDIA GeForce Line: All current and previous generation GeForce-branded parts

gitea-mirror commented

2026-05-05 03:25:39 -06:00

Author

Owner

@hebais commented on GitHub (Jun 2, 2025):

I think I found the root cause, the nvidia-smi does not work well on Jetson Platforms, here is the post:
https://forums.developer.nvidia.com/t/drivers-not-available-n-a-with-nvidia-smi-command-after-flashing-jetson-orin-nano/298011

Seems nvitop relies on nvidia-smi command to extract the GPU info, it could not get correctly on this platform.

However there is another command call "jtop" can work well, so I wonder if it is possible to add this command support as well.
The jtop has very similar UI at least, maybe the command as well seems jtop is also from nvidia.

`usage: jtop [-h] [--health] [--error-log] [--no-warnings] [--restore] [--loop] [--color-filter] [-r REFRESH] [-p PAGE] [-v]

jtop is system monitoring utility and runs on terminal

options:
-h, --help show this help message and exit
--health Status jtop and fix (default: False)
--error-log Generate a log for GitHub (default: False)
--no-warnings Do not show warnings (default: False)
--restore Reset Jetson configuration (default: False)
--loop Automatically switch page every 5s (default: False)
--color-filter Change jtop base colors, you can use also JTOP_COLOR_FILTER=True (default: False)
-r REFRESH, --refresh REFRESH
refresh interval (default: 1000)
-p PAGE, --page PAGE Open fix page (default: 1)
-v, --version show program's version number and exit`

@hebais commented on GitHub (Jun 2, 2025): I think I found the root cause, the nvidia-smi does not work well on Jetson Platforms, here is the post: https://forums.developer.nvidia.com/t/drivers-not-available-n-a-with-nvidia-smi-command-after-flashing-jetson-orin-nano/298011 Seems nvitop relies on nvidia-smi command to extract the GPU info, it could not get correctly on this platform. However there is another command call "jtop" can work well, so I wonder if it is possible to add this command support as well. The jtop has very similar UI at least, maybe the command as well seems jtop is also from nvidia. ![Image](https://github.com/user-attachments/assets/fa775b74-f365-41bd-94b2-add14cfde483) `usage: jtop [-h] [--health] [--error-log] [--no-warnings] [--restore] [--loop] [--color-filter] [-r REFRESH] [-p PAGE] [-v] jtop is system monitoring utility and runs on terminal options: -h, --help show this help message and exit --health Status jtop and fix (default: False) --error-log Generate a log for GitHub (default: False) --no-warnings Do not show warnings (default: False) --restore Reset Jetson configuration (default: False) --loop Automatically switch page every 5s (default: False) --color-filter Change jtop base colors, you can use also JTOP_COLOR_FILTER=True (default: False) -r REFRESH, --refresh REFRESH refresh interval (default: 1000) -p PAGE, --page PAGE Open fix page (default: 1) -v, --version show program's version number and exit`

gitea-mirror commented

2026-05-05 03:25:40 -06:00

Author

Owner

@hebais commented on GitHub (Jun 7, 2025):

you're right, seems jetson does not support nvml library. But I think it is still possible to use alternative apis to adapt this project

@hebais commented on GitHub (Jun 7, 2025): you're right, seems jetson does not support nvml library. But I think it is still possible to use alternative apis to adapt this project