AI Already Did It › System & Utilities › Hardware & Sensors › Hardware Monitoring

GPU-L

Name: GPU-L
Author: Administrator

GPU-Z for Linux: a GTK desktop app showing everything about your NVIDIA, AMD or Intel graphics card and monitoring its sensors in real time.

License	MIT
Generated by	Anthropic Claude Fable 5.1 —
Platforms	linux
Rating	0/100 (0 ratings) · worked for 0, didn't for 0
Downloads	0
Latest version	1.0.0 (2026-09-16)
Uploader	Administrator
Estimated generation cost	375,000 tokens (≈ $2.25) — saved so far: 0 tokens

Screenshots

GPU-L screenshot

Description

It reproduces the familiar four-tab layout:

Graphics Card – name, GPU codename and revision, process node, die size, transistor count, release date, video BIOS version with UEFI GOP flag and one-click BIOS backup, board partner and device IDs, ROPs/TMUs, PCI Express link (maximum @ current), shader count, API support, pixel and texture fillrate, memory type / bus width / size / bandwidth, driver and kernel-module signature, current and default GPU / memory / boost clocks, Resizable BAR, and Computing / Technologies checkboxes (OpenCL, CUDA, Vulkan Compute, ROCm or NVENC, Vulkan, Ray Tracing, Mesh Shaders, OpenGL).
Sensors – GPU / memory / video clocks, temperatures (GPU, hot spot, memory), fan % and RPM, memory used, GPU / memory-controller / video-engine / bus load, PCIe throughput, board power and % TDP, voltages, performance state, PerfCap reason, PCIe link generation and width, CPU temperature and clock, system memory. Every row has GPU-Z's dropdown (current / highest / lowest / average / hide), a history graph, CSV logging and reset.
Advanced – searchable key/value pages: General (OS, platform, adapter), NVIDIA (driver, clocks, P-states, clock offsets, power and temperature limits with adjustment ranges), AMD (device info, DPM tables, overdrive ranges, power/thermal, SMU metrics, firmware and IP blocks), Intel, PCI (BARs, config header, capabilities), Vulkan (driver, heaps, extensions), OpenGL (per-GPU EGL context, limits, extensions), OpenCL, CUDA, Displays (full EDID decode: native mode, size, HDR, range limits), processes using the GPU, and the video BIOS image structure.
Report – a complete dump including sensor statistics, exportable as TXT / JSON / HTML or copied to the clipboard – the local stand-in for GPU-Z's online validation.

How it works

Hardware access is done directly from Python via ctypes:

Source	Used for
NVML (`libnvidia-ml.so`)	NVIDIA name, VBIOS, CUDA cores, clocks, P-states, power/thermal limits, fans, loads, PCIe throughput, throttle reasons
amdgpu sysfs / hwmon / `DRM_IOCTL_AMDGPU_INFO`	AMD device info, DPM tables, sensors, `gpu_metrics`, firmware, unprivileged VBIOS dump
i915 / xe sysfs	Intel frequencies, power, temperature, VRAM
Vulkan, OpenCL, CUDA, EGL/OpenGL	API versions, drivers, extensions, memory heaps, limits
sysfs PCI + `pci.ids`	IDs, board partner, PCIe link, BARs / Resizable BAR
Built-in spec database	~360 NVIDIA, AMD and Intel products: process node, die size, transistors, release date, reference clocks

Clock and bandwidth conventions match GPU-Z exactly – a 21 Gbps GDDR6X RTX 4090 shows 1313 MHz memory and 1008 GB/s.

Also included

Headless CLI: --report, --json, --html, --sensors, --dump-bios
Window screenshots, settings menu (refresh rate, keep on top, log file)
.deb package, AppImage, installer script, man page, 256×256 icon
Unit tests for the database, parsers and sensor poller

Everything – research, architecture, hardware backends, the spec database, the GTK UI, packaging and tests – was produced by Claude in a single session and verified live on an RTX 4090, an RTX 5060 Ti and a Ryzen 9950X3D integrated GPU.

Downloads

gpu-l_1.0.0_all.deb — Linux, 97.2 KB
GPU-L-1.0.0-x86_64.AppImage — Linux, 1010.5 KB
gpu-z.tar.gz — Source, 606.9 KB

Original prompt(s)

Prompt 1

Create GPU-Z application equivalent for Linux.

- Do serious research before implementing any code, here is the link that may help (If you cant reach them, use chrome, I installed Claude extension already):
    - https://www.techpowerup.com/gpuz/
    - https://www.techpowerup.com/download/techpowerup-gpu-z/
    - https://en.wikipedia.org/wiki/GPU-Z
    - https://www.guru3d.com/download/gpu-z-download-techpowerup/

- It should be desktop application of course

- Do your best, good luck.

Tags: amd, gpu, gpu-z, hardware-monitoring, nvidia, overclocking, sensors, system-information, temperature, vulkan

Open in AI Already Did It