Skip to content

amin-mamandi/per-bank-dram-artifacts

Repository files navigation

RTAS 2026 Artifact

This repository contains the experimental artifact for the RTAS 2026 paper: Per-Bank Memory Bandwidth Regulation for Predictable and Performant Real-Time Systems.

Repository Layout

  • figure1/ - MLP experiments
  • figure2/ - Synthetic benchmark experiments
  • figure3/ - Real-world workload experiments
  • setupBoard.sh - Board and test-environment setup script

Requirements

  • python3
  • gnuplot
apt-get update
apt-get install gnuplot

DRAM Address Mapping

Please refer to the drama-pp README for DRAM address mapping instructions. After generating map.txt, copy it into IsolBench/bench/.

Reproducing Results

Run the setup command with sudo.

git clone --recurse-submodules https://github.com/amin-mamandi/RTAS26-Artifacts
cd RTAS26-Artifacts
sudo ./setupBoard.sh

Figure 1: MLP Experiments

Run the experiment with sudo. For the paper results, use <corun> as 0 or 3:

  • 0: one pll process
  • 3: four concurrent pll processes total
cd figure1
sudo ./mlptest.sh <maxmlp> <corun>
gnuplot plot-mlp.gp

Figure 2: Synthetic Benchmark Experiments

Run the benchmark command with sudo.

cd figure2
sudo ./run-synthetic.sh
python3 plot_synthetic_sd.py [--err]
python3 plot_synthetic_bw.py

Figure 3: Real-World Workload Experiments

Run the attack scripts with sudo. Use <benchmark> as either sdvbs or matmult.

cd figure3
sudo ./all-banks-attack.sh <benchmark>
sudo ./single-bank-attack.sh <benchmark>
python3 extract-data.py
gnuplot plot-slowdown.gp

About

IEEE RTAS 2026

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors