This archive provides the raw schedulability data for the schedulability experiments reported on in Sections V and VI of the paper:
There are seven folders to corresponding six different experimental setups.
emstada
: The main experiments using the task set generator provided by Emberson, Stafford, and Davis (see paper for details). These are discussed in Section V of the paper.
individual
: The same experimental setup as in emstada
, but including individual results for all considered heuristics. These experiments are not discussed in the paper due to space constraints.
emstada-chunk
: The same setup as in emstada
, but with the minimum chunk-size constrained imposed as discussed in Section VI.
mp
: The experiments to explore the upper bounds on preemption/migration rates as discussed in Section VI.
unc
: Experiments similar to the emstada
experiments, but using UNC-style task-set generation methods.
unc-chunk
: Like emstada-chunk
, but using the UNC task-set generators.
unc-tcount
: Like unc
experiments, but instead of varying the total utilization, the number of tasks (n) is varied.
Each folder contains files in comma-separated values (CSV) format. Each CSV file contains comments that document certain details, parameter values, and that explain its structure. In particular, the last comment line before the actual data contains an explanation of all columns.
For further details regarding the experiments and the interpretation of the data, please refer to the artifact evaluation instructions.