cuda bandwidthtest and matrixmul compile and runtime performance graphs  Scheduling frequency: 12h , nodes=1:ppn=16:xk,walltime=00:05:00 


Tests: cudatoolkit module, gpu performance