So at Super Computing 2007 we announced a new name called Windows HPC Server 2008 ( instead of Windows Compute Cluster Server -- it never really did zing - clusters what is that all about) and the fact that we are shipping the first beta of the V2 product. The second announcement is the creation of the Parallel Computing Initiative, to create a set of common multi-core development tools across desktop and cluster.
So whats in the new version - well quite a lot :-
- New high-speed networking support
- Highly efficient and scalable cluster management tools
- Support for partners’ clustered file systems
- Advanced failover support
- A new job scheduler which links our investments in distributed computing
Giovanni in the team is going to publish a screencast on TechNet EDGE on all the cool new features as well as blogging about it on his blog
On the perf side, we have been doing quite a bit of work. We have upgraded our 2,048-core test cluster to Windows HPC Server 2008, and this improved our Linpack benchmark results by 30%, which matches the very best Linux efficiencies. This is pretty cool as it is the same hardware. The other cool part is that we deployed the environment in around two hours. Linux Guys - this is huge.
We are also kicking off the Parallel Computing Initiative. So what is that all about? This will help create multi-core development tools, runtimes, programming models, libraries, and language extensions. Given that Moore’s Law will largely be driven by multi-core moving forward, its good that we are making these strategic multi-core development platform investments now.