In what has become a not-so-secret secret, Microsoft and Intel are set to unveil on March 18 their joint funding of more parallel-processing research work, according to various industry reports.
EETimes says the Wintel couple will announce it is helping fund the new Parallel Computing Lab at the University of California at Berkeley. While Microsoft officials are not publicly confirming that report, the announcement tomorrow does involve Microsoft’s ongoing interest in multicore research, a company spokeswoman confirmed.
Update on March 18: The reports were correct. Not only is UC Berkeley one of the participants in the new Parallel Computing Research Centers announced today by Microsoft and Intel, but so is the University of Illinois at Urbana-Champaign. According to the press release: “Microsoft and Intel have committed a combined $20 million to the Berkeley and UIUC research centers over the next five years. An additional $8 million will come from UIUC, and UC Berkeley has applied for $7 million in funds from a state-supported program to match industry grants.” Microsoft and Intel are billing this as the first joint industry-university collaboration “of this magnitude” that is focused on “mainstream” — not just scientific/high-end parallel-computing research.
Microsoft has a number of its own research and commercial initiatives underway in the multicore/parallel-processing arenas. The company released in late 2007 a test build of Parallel FX, which is a set of parallel extensions to Visual Studio. The Microsoft Research folks are working on the MS-ManiC (Memory Systems for Many Cores) project, which is focused on designing scalable memory-system architectures for many-core processors.
The Research team also has a request for proposal (RFP) out that will fund three-year research projects in multicore computing. The Safe and Scalable Multicore Computing RFP is for $1.5 million, which MSR anticipates making available in nine awards averageing $166,000. Proposals for that RFP were due last week, and recipients are st to be notified on April 23, 2008.
One of Microsoft’s parallel projects I’ve been following quite closely is Dryad, which is its distributed-computing infrastructure for large-scale (thousands of servers) parallel applications. Dryad is Microsoft’s answer to Google’s MapReduce technology. Dryad can scale from multicore single computers, to small clusters of computers, to data centers with thousands of computers, according to Microsoft.
On the Microsoft Research site, there’s a new research paper on Dryad which Microsoft’s researchers are set to present at the European Conference on Computer Systems (EuroSys) in Portugal next week.
Dryad is not pure research; Microsoft’s adCenter online advertising team is using the technology today. From the research paper:
“SQL Server Integration Services (SSIS) [6] supports workflow-based application programming on a single instance of SQLServer. The AdCenter team in MSN has developed a system that embeds local SSIS computations in a larger, distributed graph with communication, scheduling and fault tolerance provided by Dryad.”
The Windows Live Search team seems to be making use of Dryad, too. Again, from the paper:
“We would like to thank all the members of the Cosmos team in Windows Live Search for their support and collaboration.”
(Cosmos, as I confirmed recently, is the distributed storage layer underlying Live Search.)
One more update from March 18: Microsoft’s Tony Hey, corporate vice president of External Research at Microsoft Research, itemized during a call with press and analysts about the new industry-research collaboration several additional multicore/parallel efforts Microsoft has undertaken. These include:
- Parallel, scalable numerical libraries
- Extending Microsoft’s Robotics Studio development kit for parallel environments
- Parallel performance tools based on threads
- Transaction Memory research (in Microsoft’s Barcelona Supercomputing Center)
- Parallel programming languages (primarily in Microsoft’s Cambridge, UK, lab)
