Some benefits of exploiting partially reconfigurable devices are power consumption reduction, cost reduction, and performance improvement. To obtain these benefits, one main problem needs to be solved is the scheduling and placement.
Existing algorithms tends to allocate tasks in the positions where can be bottlenecks for future tasks to be scheduled earlier denoted as "blocking-effect". To tackle this effect, a novel 3D total contiguous surfaces (3DTCS) heuristic is proposed for equipping our algorithm with blocking-awareness.
The proposed algorithm is evaluated with both synthetic and real workloads (e.g. MDTC, matrix multiplication, hamming code, sorting, FIR, ADPCM, etc). The proposed algorithm not only has better scheduling and placement quality but also has less algorithm execution time compared to existing algorithms.
Hardware task scheduling and placement algorithms can be divided into two versions: offline and online. Offline assumes that all tasks properties (e.g. task sizes, execution time, reconfiguration time, etc) are known in advance. The offline version can t...