Dear All, There was an action on me to follow up on Jenny's comments that her team's jobs were running very very slowly at the moment, even though they are not parallel jobs and don't require GPUs. I asked Andy Richards about this and his reply was along the lines of parallel jobs have been moved to cx3 where they have been running (very inefficiently) and blocking the queues for all other users. They may have been running inefficiently but at least they do run. These jobs will move over to the new resources in the not too distant future. However, even when these jobs are moved to the new cluster we have a lack of throughput capacity. This will be looked at when the true HPC jobs have been moved to the new resource. I believe that Andy is on this list and so he should feel free to correct me if I have been inaccurate. Best, david