Research Article
Using Heuristic Value Prediction and Dynamic Task Granularity Resizing to Improve Software Speculation
Table 2
The benchmarks used in experiment.
| | Name | Package | CVAR number |
| | backprop | Rodinia 2.1 | 5 | | heartwall | Rodinia 2.1 | 2 | | kmeans | Rodinia 2.1 | 8 | | hotspot | Rodinia 2.1 | 10 | | leucocyte | Rodinia 2.1 | 16 | | srad_v1 | Rodinia 2.1 | 13 | | lavaMD | Rodinia 2.1 | 5 | | adpcm | mibench | 2 | | susan-s | mibench | 8 | | 183.equake | Spec2K INT | 5 | | 179.art | Spec2K INT | 13 | | LU | OmpSrc 2.0 | 4 | | mandelbrot | OmpSrc 2.0 | 3 | | MD | OmpSrc 2.0 | 7 | | Pi | OmpSrc 2.0 | 2 | | blackscholes | Parsec 2.1 | 0 | | badloop | self-coded | 1 | | MatMul | self-coded | 0 |
|
|
