Minnesota Supercomputing Institute purchased the first siCluster (PR on disk, finishing up and getting out tomorrow), which is a scalable storage cluster product, aimed at providing very scalable performance and capacity.
I was worried after the talk I gave at their booth. Their researcher indicated our performance wasn’t good. We had turned off some caching to avoid problems during the acceptance test for HP, the Itasca cluster vendor.
The Itasca cluster is quite nice. HP makes some good kit.
Well, I was worried that we screwed up something. I wanted to learn what, so I went to her talk.
I won’t talk about the older platform, there was some question as to whether or not the writes were being serialized by the platform.
But our storage. Should have been ~8GB/s writes. We hit 4.5-6GB/s writes for this one application. It wasn’t terrible, it compared well with the older platform. But it could have been better. Our other issue is related to the fabric. I am wondering if the combination is a problem.