PPROF(1)

📦 version 2.15, in epel 10, epel 9

💅 PPROF — the manual page, but make it make sense

okay so bestie, quick context before we dive in: pprof is part of gperftools, and it's basically the tool that reads your profiling data and tells you what your program is actually doing with its life. it's giving detective energy. let's go 🕵️‍♀️

✨ SYNOPSIS ✨

pprof [options] <program> <profile>

🌸 DESCRIPTION 🌸

pprof prints the specified cpu- or heap-profile. that's it, that's the whole pitch. you feed it a <program> and a <profile>, and it spills the tea about where your cycles (or your memory) went. lowkey iconic. 💖

🎀 OPTIONS 🎀

  • --cum — sort by cumulative data. sort of a "give me the big picture" vibe.
  • --base=<base> — subtract <base> from <profile> before display. yes, literally math-ing one profile against another. 🧮

📊 Reporting Granularity 📊

how zoomed-in do you wanna be, bestie? pick your level of detail:

  • --addresses — report at address level (maximum zoom, no cap 🔍)
  • --lines — report at source line level
  • --functions — report at function level [default] ⭐
  • --files — report at source file level

🖨️ Output type 🖨️

what format is your little report gonna come out in? choose your fighter:

  • --text — generate text report [default] ✨
  • --gv — generate Postscript and display. it'll pop it open for you, so polite 💅
  • --list=<regexp> — generate source listing of matching routines
  • --disasm=<regexp> — generate disassembly of matching routines
  • --dot — generate DOT file to stdout
  • --ps — generate Postscript to stdout
  • --pdf — generate PDF to stdout
  • --gif — generate GIF to stdout

🍓 Heap-Profile Options 🍓

these only matter when you're dealing with a heap profile, so like, vibe check the situation first:

  • --inuse_space — display in-use (mega)bytes [default] 🌸
  • --inuse_objects — display in-use objects
  • --alloc_space — display allocated (mega)bytes
  • --alloc_objects — display allocated objects
  • --show_bytes — display space in bytes (for the people who want the raw numbers, respect 🫶)
  • --drop_negative — ignore negaive differences
ok so real talk, that's literally how it's spelled in the original — negaive. not me fixing it. i would never. 💀

🦋 Call-graph Options 🦋

the fine-tuning section. this is where you get to be picky:

  • --nodecount=<n> — show at most so many nodes [default=80]
  • --nodefraction=<f> — hide nodes below <f>*total [default=.005]
  • --edgefraction=<f> — hide edges below <f>*total [default=.001]
  • --focus=<regexp> — focus on nodes matching <regexp>
  • --ignore=<regexp> — ignore nodes matching <regexp>
  • --scale=<n> — set GV scaling [default=0]

🔥 EXAMPLES 🔥

here's where it clicks, fr fr:

pprof /bin/ls ls.prof

outputs one line per procedure. simple, clean, slay.

pprof --gv /bin/ls ls.prof

displays annotated call-graph via 'gv'. pretty pictures ✨

pprof --gv --focus=Mutex /bin/ls ls.prof

restricts to code paths including a .*Mutex.* entry.

pprof --gv --focus=Mutex --ignore=string /bin/ls ls.prof

code paths including Mutex but not string. very "i want this but absolutely NOT that" of it 💅

pprof --list=getdir /bin/ls ls.prof

Dissassembly (with per-line annotations) for getdir()

pprof --disasm=getdir /bin/ls ls.prof

Dissassembly (with per-PC annotations) for getdir()

Copyright © 2005 Google Inc.

🌈 SEE ALSO 🌈

further documentation for pprof is maintained as a web page called cpu_profiler.html and is likely installed at one of the following locations:

/usr/share/gperftools/cpu_profiler.html
/usr/local/share/gperftools/cpu_profiler.html
go read those, they're the main character energy of extra docs 🧚

okay that's the whole page bestie, you're officially pprof-pilled 🎀 this was pprof (part of gperftools), dated February 2005 — vintage but still eating, ngl. mwah 💖🦋

📦 where this page comes from