Squeezing more CPU performance out of a Cray-2 by vector block scheduling | IEEE Conference Publication | IEEE Xplore