Can We Solve Linear Algebra Problems at Extreme Scale and Low Precisions?