Due to the hunger of many machine learning algorithms computing power of CPU's and GPU's are seeming to reach the limit of what they are capable of. In order to obtain more processing power alternative processing units like Field Programmable Gate Arrays (FPGA's) or software reconfigurable hardware are becoming more common in the data center. Many of the cloud providers already support these kinds of devices to be programmed by customers. So maybe it's time for a short introduction into FPGA's and show how you can program this kind of hardware.
Due to the hunger of many machine learning algorithms computing power of CPU's and GPU's are seeming to reach the limit of what they are capable of. In order to obtain more processing power alternative processing units like Field Programmable Gate Arrays (FPGA's) or software reconfigurable hardware are becoming more common in the data center. Many of the cloud providers already support these kinds of devices to be programmed by customers. So maybe it's time for a short introduction into FPGA's and show how you can program this kind of hardware. Familiarity with concepts like bits and bytes is expected. Because tooling for this kind technology requires a lot of disk space, we will be running the labs on virtual machines in the cloud. Bring a laptop with a network or wireless connection so you can create a remote desktop connection.