eXludus Offers Free Trials Of Grid Optimization Software

eXludus Technologies today announced that it is offering free 30-day trials of its new Grid Optimizer software, which is designed to be used with popular cluster workload managers including OpenPBS, PBS Pro and Torque. The free downloads for the Job Array edition of Grid Optimizer are available for up to 32 cluster nodes. “Most users have been operating for a long time with workload managers that have zero support for job arrays. With the rise of multi-core CPUs, a more dynamic approach to scheduling is quickly becoming even more important. Our new Grid Optimizer for Job Arrays provides the only powerful, full-featured implementation of these capabilities in the market place," said eXludus founder and CEO Benoît Marchand, “Users can now transparently run throughput workloads with high efficiency, achieving full utilization of their processing capacity in conjunction with their existing workload manager.” Grid Optimizer does not require replacement of current workload managers on clusters and Grids; it augments their capability and performance. It is the first tool to employ real-time dynamic schedule optimization to ensure optimal processing efficiency for any workload. Grid Optimizer's real-time virtual scheduling capability has been measured to provide 1.89 times greater throughput than PBS Pro when used alone, running a large workload of 10,000 BLAST search sequences on a 256-core cluster. Grid Optimizer has the unique ability to overlap I/O overhead behind computation. No workload management software utility or operating system multitasking process scheduler provides comparable features. Grid Optimizer also provides a profiling tool that can export timing information to Excel or produce graphical output. Information on processor utilization over time and the job iteration run time allows users and system administrators to quickly identify workload distribution requirements. With Grid Optimizer for Job Arrays, users can invoke the same executable and their existing scripts to run many related tasks or sub-jobs with varying input parameters. This is highly useful for converging on optimal solutions with throughput-oriented job streams in disciplines such as bioinformatics, financial risk and securities modeling, image rendering, and engineering analysis design optimization studies. With job arrays, sets of tasks can be run in parallel to speed solution times substantially, often by orders of magnitude. Grid Optimizer for Job Arrays is available for purchase starting at $50 per CPU for academic institutions.