Utilyze – an open source GPU monitoring tool more accurate than nvtop
The standard GPU utilization metric reported by nvidia-smi, nvtop, Weights & Biases, Amazon CloudWatch, Google Cloud Monitoring, and Azure Monitor is highly misleading. It reports the fraction of time that any kernel is running on the GPU, which means a GPU can report 100% utilization even if only a small portion of its compute capacity is actually being used. In practice, we've seen workloads with ~1–10% real compute throughput while dashboards show 100%. This becomes a problem when teams rely on that metric for capacity planning or optimization decisions, it can make underutilized systems…
In plain words
Utilyze is an open-source GPU monitoring tool that measures actual compute and memory throughput instead of relying on standard metrics like those from nvidia-smi or nvtop. It samples hardware performance counters and reports utilization as a percentage of theoretical hardware limits, giving a more accurate picture of GPU capacity. The tool is designed for teams making capacity planning and optimization decisions who need to distinguish between GPUs that appear fully utilized on traditional dashboards but are actually underperforming.
written from the facts on this page · September 2026
From the sources
In the maker’s words, at launch
The standard GPU utilization metric reported by nvidia-smi, nvtop, Weights & Biases, Amazon CloudWatch, Google Cloud Monitoring, and Azure Monitor is highly misleading. It reports the fraction of time that any kernel is running on the GPU, which means a GPU can report 100% utilization even if only a small portion of its compute capacity is actually being used. In practice, we've seen workloads with ~1–10% real compute throughput while dashboards show 100%. This becomes a problem when teams rely on that metric for capacity planning or optimization decisions, it can make underutilized systems look saturated. We're releasing an open-source (Apache 2.0) tool, Utilyze, to measure GPU utilization differently. It samples hardware performance counters and reports compute and memory throughput relative to the hardware's theoretical limits. It also estimates an attainable utilization ceiling for a given workload. GitHub link: https://github.com/systalyze/utilyze We'd love to hear your thoughts!
Does a similar job
all alternatives →- SASping – An HTTP/TCP latency tool that's easy on the eye2025 · dseltzer.gitlab.io · ▲178
I've frequently found myself using [nvitop](https://github.com/XuehaiPan/nvitop) to diagnose GPU/CPU contention issues. The two best things about it are: - It's easy to install if I can access pip in the container - It makes a compelling screenshot (which helps me communicate with coworkers.) With those two lessons in mind: Here is Sping! Purpose: Help observe and diagnose latency issues at layer 4+ (TCP/HTTP/HTTPS) Two good things about it: - It's easy to install if you have pip. (Available at…
- AOAn open source performance monitoring tool2024 · github.com · ▲135
Hey HN. We’re Jay and Vadim from Highlight.io (https://highlight.io). We’re building an open source [1] monitoring platform for web applications. Today we’re excited to be sharing a performance tool we’ve been working on, which helps you inspect the latency of code execution from the client to the server. As engineers at past startups, we often had to debug slow queries, poor load times, inconsistent errors, etc... While tools like Jaegar [2] helped us inspect server-side performance, we had no way to tie user events to the traces we were inspecting. In other words, although we had…
cuTile Rust: Safe, data-race-free GPU kernels in RustJun 2026 · github.com · ▲106cuTile Rust provides a safe, tile-based kernel programming DSL for the Rust programming language. It features a safe host-side API for passing tensors to asynchronously executed kernel functions. - NVlabs/cutile-rs
L9gpu – GPU telemetry that ties each GPU to the K8s pod or Slurm jobJul 2026 · github.com · ▲20GPU Observability with workload attribution. One OTLP agent per node ties hardware metrics (NVIDIA, AMD, Intel Gaudi) to the K8s pod or Slurm job burning the GPU. - last9/gpu-telemetry
- IMI made a calculator to show cost savings of serverless GPUs vs. AWS2022 · paylesstoaws.com · ▲25
- IWI wrote a tool in Rust for tracking all allocations in a Linux process2023 · github.com · ▲113
More dev tools this month
the category →
OpenTrailPaper is open-source bike computer firmware for the LilyGO T5S3 4.7" E-Paper PRO. It supports offline maps, GPX routes, FIT recording and Bluetooth sensors.
Dev tools · 2d ago · opentrailpaper.com
Open-source GTM skills for technical founders
Dev tools · 30d ago · gtmcofounder.com
Launched alongside, April 2026
the whole month →
- AG
Thought the resources for GPU arch were lacking, so here we are
Life & fun · Apr 2026 · jaso1024.com
- IB
Built a ~9M param LLM from scratch to understand how they actually work. Vanilla transformer, 60K synthetic conversations, ~130 lines of PyTorch. Trains in 5 min on a free Colab T4. The fish thinks the meaning of life is food. Fork it and swap the personality for your own character.
AI · Apr 2026 · github.com
- BC
Life & fun · Apr 2026 · sam-burns.com
- IB
With social media and now AI, its important to keep the indie web alive. There are many people who write frequently. Blogosphere tries to highlight them by fetching the recent posts from personal blogs across many categories. There are two versions: Minimal (HN-inspired, fast, static): https://text.blogosphere.app/ Non-minimal: https://blogosphere.app/ If you don't find your blog (or your favorite ones), please add them. I will review and approve it.
AI · Apr 2026 · text.blogosphere.app