[zifpanachr23]

Yeah the architectures focus on different things. The huge caches on the z series chips are designed primarily around the kind of latency sensitive workloads more common in finance than the massive floating point throughput often needed for big scientific computing.