Skip to main content

Table 9 Difference of simulation and proposed model under multi-threaded workloads for a 16-core CMP

From: A novel power model for future heterogeneous 3D chip-multiprocessors in the dark silicon age

Workload CMP Cores (%) Cache hierarchy (%) NoC (%)
blackscholes Baseline 4.225 0.714 6.452
Hybrid 3.077 3.279 1.754
bodytrack Baseline 4.348 4.412 3.226
Hybrid 3.077 3.636 3.448
canneal Baseline 5.000 9.091 3.333
Hybrid 4.762 14.151 0.000
dedup Baseline 6.977 1.282 0.000
Hybrid 7.692 3.846 0.000
facesim Baseline 3.077 3.077 3.333
Hybrid 1.639 3.704 3.571
ferret Baseline 1.991 0.915 3.333
Hybrid 1.613 3.774 3.448
swaptions Baseline 4.360 4.425 3.226
Hybrid 2.326 1.852 0.000
fluidanimate Baseline 4.878 6.410 3.333
Hybrid 5.263 5.455 3.448
freqmine Baseline 2.985 2.941 3.333
Hybrid 1.563 3.704 3.333
vips Baseline 5.172 16.923 3.333
Hybrid 1.818 1.852 3.448
x264 Baseline 1.695 2.899 3.448
Hybrid 1.754 1.852 0.000
Average Baseline 3.726 4.424 3.029
Hybrid 3.144 4.282 2.041