This repository contains the modified gem5 source code for collecting microarchitectural statistics of a core in chiplet-based multi-core systems. It generates runtime statistics for dynamic micro instructions (dyn inst), the Reorder Buffer (ROB), and the Load Queue (LQ). For further details, please refer to the following paper.
The Fake-Busy and True-Idle Problems of Running Graph Applications on Chiplet-Based Multi-cores.
Rashid Aligholipour, Yuan Yao,
IEEE International Symposium on Workload Characterization (IISWC)-2025
To compile, the following command should be executed:
scons PROTOCOL=MESI_Two_Level build/X86/gem5.fast -j 16 --linker=moldIt should be noted that linker is optional. Other linker can be used or even in can be omiited.
There are two scripts for running the simualator. One of them for spec-2017 and another one for GAPBS. Both are in the artifact directory.
cd artifact/then with following command they can be executed.
python3 run_spec.pypython3 run_gapbs.pyThis repo for executing, it requires,
- Binaies of the programs (it must compiled statically).
- Checkpoint of Region-of-Interest (ROI).
- Datasets/Inputs
Due to big size of of theses files, it must be provided by users. Once all theses files are collected it would be ready for execution.