enelog

enelog is a lightweight command-line utility for measuring CPU power consumption using Intel's RAPL (Running Average Power Limit) interface. It is implemented in C and built with CMake.

Features

• CPU package power & energy via Intel RAPL
• Optional DRAM domain power & energy (-d)
• Optional GPU total and per-GPU power & energy via NVIDIA NVML (-g, -G)

Requirements

• OS: Linux with /sys/class/powercap (RAPL) support. You may need to have root privileges.
• OS: Linux with /sys/class/powercap (RAPL) support. You may need to have root privileges.
• CPU: Intel processor with RAPL package—and optionally DRAM—domains
• GPU (optional): NVIDIA driver + NVML (libnvidia-ml.so) if using -g/-G
• Permissions: Reading energy_uj may require sudo on some systems

Build Instructions

Make sure you have CMake (version 3.22 or higher) and a C compiler installed.

git clone https://github.com/your-username/enelog.git
cd enelog
cmake -B build
cmake --build build

After building, the executable will be available at build/enelog.

Building

• CPU-only

cmake -B build \
  -DCMAKE_BUILD_TYPE=Release \
  -DCMAKE_C_FLAGS="-I/usr/local/cuda/include" \
  -DCMAKE_EXE_LINKER_FLAGS="-L/usr/lib/x86_64-linux-gnu -lnvidia-ml"
cmake --build build

Adjust include/lib paths for your distribution (e.g., /usr/lib, /lib/x86_64-linux-gnu, /usr/local/cuda/lib64).

• NVML-enabled build (includ GPU)

cmake -B build -DUSE_NVML=ON -DCMAKE_BUILD_TYPE=Release
cmake --build build

Usage

./enelog [-i <interval>] [-t <timeout>] [-d] [-D] [-E] [-H] [-h] [-g] [-G]

• -i <interval>: Interval in seconds between power measurements (default: 1 second)
• -t <timeout>: Timeout in seconds after which the program exits (default: 120 seconds)
• -d: Enable DRAM power/energy (if DRAM RAPL domain exists)
• -D: Prepend MM-DD to the timestamp (otherwise HH:MM:SS)
• -E: Include energy(J) fields in output (otherwise power only)
• -H: Print a one-line header describing the output columns
• -h: Show help and exit
• -h: Show help and exit

GPU (requires NVML and a build with USE_NVML=ON):

• -g: Enable GPU power/energy measurement via NVML
• -G: Also print per-GPU metrics(implies -g)

Output Schema

Timestamp

• Default: HH:MM:SS
• With -D: MM-DD HH:MM:SS

Columns (by flags)

• Base (CPU only, no -E):

  [time] CPU(W)
  

• CPU with energy (-E):

  [time] CPU(W) CPU(J)
  

• CPU + DRAM (-d, no -E):

  [time] CPU(W) DRAM(W)
  

• CPU + DRAM with energy (-d -E):

  [time] CPU(W) CPU(J) DRAM(W) DRAM(J)
  

• Add GPU total (-g):

  [time] ... GPU(W) [GPU(J) if -E]
  

• Add per-GPU (-G):

  [time] ... GPU(W) [GPU(J)] GPU00(W) [GPU00(J)] GPU01(W) [GPU01(J)] ...
  

License

This project is licensed under the MIT License. See the LICENSE file for details.

Notes

• Platform support

  • Requires Linux with RAPL exposed under /sys/class/powercap.
  • Works on Intel CPUs that support RAPL. AMD/others are not supported.
  • Only package 0 (intel-rapl:0) is read; multi-socket systems beyond pkg0 are not measured.
  • The DRAM RAPL domain may not be present on all platforms.

• Permissions

  • Reading energy_uj may require elevated privileges. Use sudo or adjust permissions via a udev rule (distro-specific).

• DRAM flag (-d)

  • If your system lacks a DRAM RAPL domain, using -d will fail. Omit -d in that case.

• GPU measurement (-g, -G)

  • Requires NVIDIA driver + NVML (libnvidia-ml.so and a build with -DUSE_NVML=ON).
  • -G includes per-GPU metrics and implies -g.
  • If NVML isn’t available or no NVIDIA GPUs are present, GPU measurement will not work.

• Timing & alignment

  • Using absolute time scheduling, sampling is aligned to second-of-minute boundaries.
  • By this, timing alignment is consistent for running

• Units & fields

  • Power is in watts (W) and energy in joules (J).
  • Use -E to include energy columns; -H prints a header row; -D adds MM-DD to the timestamp.

• Accuracy considerations

  • CPU energy uses RAPL counters (platform-estimated). GPU energy is trapezoid-integrated from instantaneous NVML power.
  • Shorter intervals generally improve integration accuracy but increase overhead.

• Containers/VMs

  • Some containers or virtual machines do not expose /sys/class/powercap or NVML; measurements may be unavailable there.

• More info

  • See the Linux powercap documentation: https://www.kernel.org/doc/Documentation/power/powercap/powercap.txt