Although modern accelerator devices, such as vector engines and SmartNICs, are equipped with general purpose CPUs, access to the storage needs the mediation of the host kernel and CPUs, resulting in latency and throughput penalties. In this paper, we explore the case for direct storage access inside the accelerator applications, and discuss the problem, design options and benefits of this architecture. We demonstrate that our architecture can improve throughputs of LevelDB by 12–89%, and reduce the execution time by 33–46 % in a bioinformatics application in comparison to the baseline where the host system mediates the storage accesses.
|Title of host publication||Proceedings of the 12th USENIX Workshop on Hot Topics in Storage and File Systems (HotStorage 20)|
|Number of pages||7|
|Publication status||Published - 13 Jul 2020|
|Event||12th USENIX Workshop on Hot Topics in Storage and File Systems (HotStorage 20) - virtual|
Duration: 13 Jul 2020 → 14 Jul 2020
|Workshop||12th USENIX Workshop on Hot Topics in Storage and File Systems (HotStorage 20)|
|Period||13/07/20 → 14/07/20|