Profile memory usage

Profile memory usage in C++ — with the four heights of help laid out: do it now, make it easier for the next person to accept, work out the right move when you are stuck, and learn the pattern so it stops coming back.

4prompt heights
Open it in the interactive atlas →

The four heights

The same task, four distances: today's deadline, the next reviewer, the stuck moment, the pattern.

Execute — do the immediate task

+
I need a short memory profile on the new asset pipeline. Run a tracing session on the Windows build…
I need a short memory profile on the new asset pipeline. Run a tracing session on the Windows build loading the city map, capture peak and steady-state numbers, and send the report to Lara in tools and to me. Include the top ten allocation sites and a leak check; confirm by Friday that no single allocation exceeds 16MB during load.

Improve — make it easier to accept

+
Before I give this to Lara in tools, make the memory report actionable — highlight the three…
Before I give this to Lara in tools, make the memory report actionable — highlight the three largest persistent allocations, collapse transient spikes into a simple timeline, mark allocations from third-party libraries, and flag anything that would make an auditor hesitate to ship the build.

Decide — diagnose the stuck moment

+
The city map load shows a sudden 1.2GB peak over baseline. I captured an allocation trace but I…

Peak memory is 1.2GB higher than baseline on the city map load

The city map load shows a sudden 1.2GB peak over baseline. I captured an allocation trace but I can’t tell if the spike is caused by temporary buffering, a leaked container, or a big texture decode. I’m afraid to start freeing things blindly. What’s the likeliest diagnosis and the safest next probe to run that isolates temporary from persistent allocations?

Become — change the pattern

+
Every milestone we add features and memory usage quietly rises; we only notice before release. I…

Memory keeps creeping up between milestones

Every milestone we add features and memory usage quietly rises; we only notice before release. I spend time firefighting with ad-hoc fixes that later reappear. Which habit should I adopt to keep steady memory budgets across sprints, and what lightweight checks to add to CI so increases are caught early?

Next to this one

Other programming language work people do in C++.

Every task here came from the work, not from a feature list — which is why the prompts name what you want done and never the button that does it. The tool changes; the work does not.
Copyright © LLOS.ai · 2026 — original pedagogy, voice, and design — all rights reserved.

The rest of the map

Same library, five ways in.