What people are saying - Write a review
We haven't found any reviews in the usual places.
High Performance Numerical Libraries
Architectural Resource Management
A Performance Study of Outoforder Vector Architectures and Short Registers
27 other sections not shown
algorithm applications architecture bandwidth benchmark block branch branch prediction buffer cache cluster communication compiler copies cycle data dependence data-flow analysis data-flow values distributed dynamic enddo evaluate execution fetch Figure forking Fortran function global grid hardware High Performance Fortran IEEE implementation infinity Lambda test instructions integer interface iteration space latency linear load loop fusion loop nest machine matrix MBCF metacomputer method misprediction multipath multiple multiprocessor Myrinet node number of processors operations optimization overhead parallel computer parallel programs partition path pattern placement prediction predictor prefetching problem Proc protocol queue radix sort random number reduce references requires reuse run-time scalar scheduling scheme sequential servers Shmem simulation solver speculative speculative execution speedup SPMD statements stride structure superscalar synchronization Table task techniques texture mapping thread tiling tion transformation variable vector workstations