Here is a collection of highlights from this week’s news stream as reported by HPCwire.

Pico Computing Accelerates Cracking of NTLM Authentication Protocol by 500X

First African School of Physics Announced

Purdue University Increases Research Capabilities with HP Datacenter

Industry Parallel Storage Experts Form New Company

Penguin Computing’s ‘POD’ HPC On-Demand Platform Hosts RealityServer Platform

Researchers Receive NSF Award to Encourage International Science Collaboration

DOE Awards Argonne Projects 200 Million Hours of Supercomputer Time

Data Sorting World Record Falls as UCSD Computer Scientists Break Terabyte Barrier

Intel Confirms Light Beams Can Replace Electronic Signals for Future Computers

Moab Brings Automation to Amazon EC2 Cluster Compute Instances

First Results from Large Hadron Collider Announced

Pico Computing Unveils FPGA Library for Signal Processing and Video Analytics

NIWA Supercomputer Switched On

Cray XE6 Officially Launched

This week Cray announced the delivery of the first full-size XE6 system (formerly codenamed “Baker”), fulfilling its Q3 production launch deadline. This is an important milestone for Cray, as it marks the first shipment of a production-ready Cray XE6 supercomputer — the first in a line of such systems that will be shipped to customers over the next few months.

Aside from saying that it is a multi-cabinet system, Cray did not elaborate on the size of the supercomputer nor did it disclose the name of the customer. Cray did say that it had sent a beta machine to the the Swiss National supercomputing centre (CSCS) in June, in addition to having shipped a number of small, test and development systems to other customers.

The 20-blade, single-cabinet machine that was delivered to CSCS uses the latest  2.1GHz, 12-core AMD Opteron (Magny-Cours) CPUs and sports 160 compute sockets for a total of 1,920 cores. The system, named Piz Palu, has a theoretical peak performance of 16 teraflops and 2.5 terabytes of memory.

Peter Ungaro, president and CEO of Cray, thanked the Swiss National Supercomputer Center for its pivotal role in preparing the Cray XE6 supercomputer for full production status. As a beta-testing partner, CSCS and its user community were able test Cray’s latest hardware and software technologies and were thus granted early familiarization with the system. As such, the Cray XE6 system is part of a joint collaboration between Cray and CSCS.

The Cray XE6 includes the much-heralded Gemini interconnect network, with its promise of increased performance and greater fault tolerance over the previous SeaStar technology. The Gemini interconnect also offers better support for Partitioned Global Address Space (PGAS) languages, such as Co-array Fortran (CAF) and Unified Parallel C (UPC). Other enhancements to the XE6 are improved network resiliency, a mature scalable software ecosystem and the latest version of the Cray Linux Environment. The XE6 is fully upgradeable from a Cray XT5 or Cray XT6 system.

Cray has announced a number of customer wins over the last year for its XE6 line. They include:

• Korea Meteorological Administration
• DOE’s National Energy Research Scientific Computing Center (NERSC)
• U.S. Air Force Research Laboratory
• Arctic Region Supercomputing Center
• U.S Army Engineer Research and Development Center
• National Nuclear Security Administration (in a joint partnership with Los Alamos and Sandia National Laboratories)
• Engineering and Physical Sciences Research Council (EPSRC) (which is the managing agent for the High-End Computing Terascale Resource (HECToR) project located at Scotland’s University of Edinburgh)
• National Oceanic and Atmospheric Administration (through a partnership with the Oak Ridge National Laboratory)

As of the May 2010 Cray XE6 “pre-launch,” Ungaro revealed the company had already secured more than $200 million in contracts, which brings us to the following paragraph that appears in this week’s announcment:

Although Cray has begun shipping Cray XE6 systems, to obtain revenue and cash from these sales, or any future Cray XE6 deliveries, the company must obtain customer acceptances of the systems typically based on a multi-week process of performance, functionality and reliability testing.

In other words: Cray can only book the revenue once the machine is officially accepted, which can take a while. I can only surmise the financial legalese is meant to temper investor expectations concerning fiscal year projections.

A Lotta SIGGRAPH (GPU-related) News

There was a lot of graphics-related news coming out of the SIGGRAPH conference this week. SIGGRAPH, which stands for “Special Interest Group on Graphics and Interactive Techniques,” is the arm of the Association for Computing Machinery (ACM) dedicated to promoting computer graphics and interactive techniques.

There seemed to be a bit more HPC-related news out of SIGGRAPH than in previous years — perhaps this has something to do with the GPU’s growing popularity, and the excitement over the “Fermi” architecture. Which brings us to NVIDIA’s announcement of its Fermi-class Quadro graphics processing units, suited to creating the “computational visualization workstation.”

From that announcement:

The NVIDIA Quadro Plex 7000 array, and Quadro 6000, Quadro 5000 and Quadro 4000 GPUs feature the new NVIDIA Scalable Geometry Engines and leverage NVIDIA Application Acceleration Engines (AXE) to enable the world’s fastest performance across a broad range of CAD, DCC and visualization applications. Rated at an unheard of 1.3 billion triangles per second in raw performance, the Quadro 6000 enables users to interactively work with models and scenes that are five times more complex than ever before.

This Quaddro professional graphics solution now offers Error Correction Codes (ECC) memory and double precision floating point performance to enable high-end applications that require the utmost accuracy.

Companies offering the Fermi-based Quadro solution include Dell, HP, Lenovo, BOXX Technologies, NextComputing, and AMAX. Distributors include PNY Technologies in North America and Europe, Leadtek in Asia Pacific, and ELSA in Japan.

NVIDIA also released its Application Acceleration Engines (AXE) at SIGGRAPH. The Engines are optimized for the newest Quadro graphics processing units (GPUs) based on the company’s Fermi architecture. Last in the NVIDIA news triad is the NVIDIA 3D Vision Pro, a new 3D stereoscopic solution that allows digital artists, product designers, and physicians to view their work in 3D. NVIDIA has been developing the technology since 1999.

The next newsworthy story also involves, you guessed it, NVIDIA. PEER 1 Hosting announced the largest, public hosted GPU cloud, which runs the RealityServer 3D Web application, developed by mental images, an NVIDIA company. The RealityServer platform incorporates NVIDIA Tesla GPUs and 3D Web services software to provide applications over the Web. For more on this announcement, read our feature story.

Notable announcements were also made by BlueArc, which highlighted its role in providing storage solutions to the media and entertainment industry, and Khronos, which released its OpenGL 4.1 specification.