Hybrid CPU–GPU implementation of the transformed spatial domain channel estimation algorithm for mmWave MIMO systems

Abstract Hybrid platforms combining multicore central processing units (CPU) with many-core hardware accelerators such as graphic (GPU) can be smartly exploited to provide efficient parallel implementations of wireless communication algorithms for Fifth Generation (5G) and beyond systems. Massive multiple-input multiple-output (MIMO) systems are a key element the 5G standard, involving several tens or hundreds antenna elements communication. Such high number antennas has direct impact on computational complexity some MIMO signal algorithms. In this work, we focus channel estimation stage. particular, develop implementation recently proposed algorithm. Its performance in terms execution time is evaluated both CPU GPU. The results show that computation blocks algorithm more suitable implementation, whereas other parts efficiently implemented GPU, indicating hybrid CPU–GPU would achieve best practical applications based tested platform.