hw-smi - a minimal, cross-compatible CPU/GPU telemetry monitor @ProjectPhysX

Uploaded April 2026 | Updated September 2026, 1 week ago

It was a long journey, now my fun side project is done and #opensourced! I present:
hw-smi - github.com/ProjectPhysX/hw-smi 🖖🧐
A minimal, cross-compatible #CPU/#GPU telemetry monitor with accurate data directly from vendor APIs and beautiful #ASCII visualization.

How much #VRAM bandwidth does an application or a game pull? Is the traffic over #PCIe a bottleneck? What's the CPU/GPU load, RAM/VRAM occupation, temperatures, power draw, clock frequencies? hw-smi works with all CPUs and all #Nvidia/#AMD/#Intel GPUs, on both Windows and #Linux.

There are lots of cool hardware monitoring tools already - but they all are either OS-specific, only for CPU or GPU, only for one particular vendor's GPUs, and none of them tell you the VRAM/PCIe bandwidth to give clues about application bottlenecks. 2 years ago I thought to myself: I can do this better. Hardest part was testing the vendor APIs, every single counter, and I needed hardware for that. That was fun, I found & reported lots of bugs. Accordingly I had to do extensive hardening against broken API counters, entire chains of fallbacks just to get one value reliably. Testing, testing, testing!

But it was all worth it. Now I share this powerful tool with the world, for free. Have fun!

Video shows hw-smi on my Frankenstein Nvidia Titan Xp + AMD Radeon RX 7700 XT + Intel Arc B580 triple-GPU test rig, loaded with a #FluidX3D multi-GPU #CFD simulation on all 3 GPUs. Look at all the counters go brrr!

hw-smi - a minimal, cross-compatible CPU/GPU telemetry monitor

PhysX3D [GRAVITY SIMULATION] Extreme Performance

$From Boxes to photorealistic Fluid (marching-cubes + raytracing) From Boxes to photorealistic Fluid - how does ot work? The fluid simulation is done with the lattice Boltzmann method (LBM) with the Volume-of-Fluid (VoF) extension and piecewise linear interface construction (PLIC) for surface tension. VoF-LBM: https://www.researchgate.net/publication/349552856_High_Performance_Free_Surface_LBM_on_GPUs PLIC: https://www.mdpi.com/2079-3197/10/2/21 VoF-LBM provides a 3D grid with fluid fill levels, used as isovalues for marching-cubes, which generates a contiguous triangle mesh representing the water surface. Marching-cubes: http://paulbourke.net/geometry/polygonise/ With raytracing I shoot light rays from the camera through the 3D isogrid. For each grid cell that a ray traverses, I run marching-cubes to generate the triangles on-the-fly, then check for ray-triangle intersections with any of these triangles. If the ray intersects, I dont use the flat-face-surface nromal of the triangle, as this would lead to visual artefacts on the reflected/refracted rays. Instead, I interpolate the surface normal at the intersection point in order to have a smoothly appearing surface. How does surface normal interpolation work with marching-cubes? I get the normals directly at the lattice points with gradient (central derivative) on the isogrid, then linearly interpolate to get the normals at triangle vertices and linearly interpolate again to where the ray intersects the triangle with Barycentric coordinates on the triangle. The 3D lattice that the LBM provides already is one of the ideal acceleration structures for raytracing. Any ray only has to check for possible intersections with iso-surface triangles contained within the traversed grid cells. This makes it - when the lattice is not too large - even run in real time on non-RTX GPUs. New video with surface normal interpolation: https://youtu.be/UUYrWUpPQwY Old video without surface normal interpolation: https://youtu.be/9Rfu9ZqW894 The simulation setup you see here is a physically accurate 4mm diameter terminal velocity raindrop impact in water, although at rather low 256x256x218 resolution for this real time demo. See my published research paper in Microplastics and Nanoplastics, where the simulation is validated to great extent: https://microplastics.springeropen.com/articles/10.1186/s43591-021-00018-8 This simulation as well as the raytracing graphics are running in real time on a single Nvidia Titan Xp GPU. The simulation software used is FluidX3D, an OpenCL implementation of the lattice Boltzmann method. FluidX3D runs at peak hardware efficiency on the worlds fastest data-center GPUs (8799/5232 MLUPs/s on the Nvidia A100/V100 @ D3Q19 SRT FP32), on gaming GPUs and CPUs and it even runs my smartphone. For more performance details, see: https://arxiv.org/pdf/2112.08926.pdf #FluidX3D #OpenCL #Raytracing #GPU #LBM$

How aerodynamic is Boeing 767-400ER? (return from SC24, landing in Munich)

Formula 1 Ferrari SF71H in FluidX3D CFD at 10 Billion Cells with Wheel Rotation - 71TB CFD Data

FluidX3D v2.7 Update: New Slice and Surface Pressure Data Visualization Modes

Smallest CRT monitor in the world with a Raspberry Pi

Airodynamics of an arm chair on 4x A100 GPUs

OpenCL GPU Programming for HPC Applications (ChEESE Webinar - Center of Excellence in Solid Earth)

USS Enterprise CFD simulation on my gaming PC