Ask in plain English
“How can I speed up my app?” “Why is there so much GPU idle time?” Seer’s AI interprets your question, then builds a custom analysis pipeline to answer it.
AI-powered HPC Performance Analysis
A new way to understand your code. Seer launches targeted analysis pipelines to pinpoint performance issues in your app.
01/Product
Straight to the problem.
Stop sifting through mountains of data and complex traces. Seer finds the bottleneck for you.
Evidence-backed results
Every analysis pipeline produces annotated visualizations and findings, so you always see where conclusions come from.
Zero-friction adoption
Seer captures application data without any modifications to your source code. And our web-based interface means no new software to install.
02/How it works
From question to answer in three steps.
Seer simplifies the profiling process so you can spend time actually writing code.
Collect
Attach Seer’s collector to your run. Minimal overhead, no source changes or recompilation required.
Ask
Describe what you want to know in plain English. Seer’s AI turns the question into a concrete analysis plan.
Explore
Answers land in a Vista with interactive visualizations you can drill into, annotate, and share with your team.
03/Vistas
Take in the view.
A Vista is your customizable analysis dashboard for a single run, composed of interactive visualizations built specifically for HPC data.
Overview first
See the whole run at a glance.
Start with high-level views of execution timelines and node heatmaps. Seer highlights issues so you don't need to sift through mountains of data on your own.
Drill down
Zoom straight to the call that matters.
Click into any hotspot to expand function-level calls, per-rank traces, and the supporting analyses. Annotations from Seer’s AI explain what it found and why it flagged it.
04/Collaborators wanted
Can your app run faster?
Of course it can! We’re looking for teams with real codes and real performance
problems.
Fill out the form below to get in touch with us.
Research groups
University and lab teams exploring parallel performance on new architectures.
Production HPC
Application owners running at scale who need answers faster than a traditional profiler gives them.
Facility staff
Center and system teams supporting many users across many stacks.
Seer is funded by the U.S. Department of Energy, Office of Science, Office of Advanced Scientific Computing Research, through the Small Business Innovation Research (SBIR) program. The WorkVisualizer prototype was developed under SBIR Phase I Award DE-SC-0024832.