I was poking around because a customer was inquiring about GPU and parallel computing. What I found out is CUDA from nvidia (AMD has one too which will be another post). This is basically C compiler made for the teraflop computing power of Nvidia’s many cored GPUs.
Poking around their website, the only one that works with rack mounted servers (Dell Poweredge 1950 3rd gen with PCI express riser card, Dell R900, HP Proliant 160 G5, and a bunch of Sun x22__ and x4___ ) is a 1U server Tesla S1070 connected by dual PCI Express 2.0 cable connections. Was hoping for an internal rack mounted solution but beggers can’t be choosers!