4 ms·3x Speedup and 2x Lower TCO for Apache Spark ML Using Amazon F1s with FPGAs10 points by bugsense 8y ago