Add memory benchmark with valgrind massif

[Copilot]

Adds make benchmark-mem to profile ART memory overhead using valgrind massif, testing uniform and zipfian key distributions with nullptr values.

Implementation

• bench/memory.cpp: Standalone executable accepting n_keys and distribution type (uniform|zipf)
• Makefile: benchmark-mem target runs massif on 100k keys for both distributions, outputs reports
• CMakeLists.txt: New bench-mem executable linked against art and zipf libraries

Usage

# Run default benchmarks (100k keys, both distributions)
make benchmark-mem

# Custom run
valgrind --tool=massif --massif-out-file=massif.out ./build/bench-mem 250000 zipf < /dev/null
ms_print massif.out

Massif output shows peak heap usage and allocation breakdown by ART node types (node_4, node_16, node_48, node_256).

Warning

Firewall rules blocked me from connecting to one or more addresses (expand for details)

I tried to connect to the following addresses, but was blocked by firewall rules:

• esm.ubuntu.com
  • Triggering command: /usr/lib/apt/methods/https /usr/lib/apt/methods/https (dns block)

If you need me to access, download, or install something from one of these locations, you can either:

• Configure Actions setup steps to set up my environment, which run before the firewall is enabled
• Add the appropriate URLs or hosts to the custom allowlist in this repository's Copilot coding agent settings (admins only)

Original prompt

---

This section details on the original issue you should resolve

<issue_title>benchmark missing space / mem used</issue_title>
<issue_description>possible to include the benchmark of space / mem used?</issue_description>

Comments on the Issue (you are @copilot in this section)

@rafaelkallis hi @hiqsociety, would a `make benchmark-mem` command solve your needs? The command would run a file using [valgrind massif](https://valgrind.org/docs/manual/ms-manual.html). For the benchmark itself I would test two scenarios of interest: 1) uniformly distributed keys and 2) zipfian keys. I would fill the tree with `nullptr` values to only measure data structure overhead.

Would that solve your use case? Do you have any other parameter (dependent variable) wishes for the benchmark?</comment_new>

• Fixes benchmark missing space / mem used #16

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.