The Open Group quest for parallelism

Yet another round of financing for academic research in parallelism, this time in Europe. EE Times reports:

Supported by the European Commission, the Java Environment for Parallel
Real-time Development (JEOPARD) project is investing over €3.3 million
(about $ 4.9 million) an advanced framework
for real-time Java running on multicore and parallel systems.

Led by The Open Group (Reading, England), the JEOPARD consortium
includes four universities and research institutes: University of York
(England), Vienna University of Technology (Austria), FZI (Research
Center for Information Technologies at the University of Karlsruhe,
Germany) and the Technical University of Cluj-Napoca (Romania); three
industrial manufacturers: EADS NV (Schiphol-Rijk, Netherlands),
RadioLabs (Rome, Italy) and SkySoft (Lisbon, Portugal); and two
embedded systems technology suppliers: Aicas (Karlsruhe, Germany) and
Sysgo (Klein-Winternheim, Germany).




Technorati Tags: grid computing, parallelism

Berkley lab

EE Times has an article on a $10M Intel and Microsoft grant that will finance the Berkley lab, a group focused on research in parallel programming. The move is meant to close the gap between hardware multi-core architectures, now in mature state, and software development tools.

The original Berkley lab white paper was published in December 2006 and is available here.

Powered by ScribeFire.

Technorati Tags: parallel programming, distributed computing

Another Sun open source acquisition

VentureBeat reports that Sun has acquired Innotek, a company developing an open source desktop virtualization software called VirtualBox. Sun move follows the acquisition of MySQL and other open source companies such as StarOffice.

There are now at least three different offers to run virtual machines: Parallels, VMWare and VirtualBox. This acquisition marks the definitive commoditization of computer hardware. Today you can run any of the major OSes (MacOS, Windows, Linux) on the two major consumer machine: Mac and PC hardware are in facts completely compatible (through Bootcamp or a virtual machine product) and Intel-compatible Linux servers are replacing older HP and Sparc hardware in the corporate world.

One can't say proprietary architectures have been very successful over the last 10 years. DEC (acquired by Compaq), HP, IBM, SGI, Sun and Apple have all converged towards an Intel-compatible platform. Ironically, this represents the best time for consumers to consider alternatives to Windows-only machines. Microsoft is now put under pressure from both the convergence towards a common platform (thus making the OS less relevant) and the Internet. The old proprietary OS model is fading out, as are other parts of the enterprise infrastructure.


Technorati Tags: open source

Powered by ScribeFire.

Multicore virtual platforms

I had never heard of multicore virtual platforms before. This article from EE Times claims multi-core debugging issues can be resolved with the use of virtual platforms:

Even a parallel compiler does not address all the challenges posed when processors overwrite each other's memory areas, lock each other out of shared memory and race to complete tasks sooner (or later) than expected. The debugging capabilities of a physical prototype can typically handle freezing only one processor at a time, leaving the others running, and likely writing over memory locations needed for debugging.

Virtual platforms overcome this challenge by freezing all the processors simultaneously (including all IPC hardware and peripheral clocks) at breakpoints, giving complete visibility into all system states. To help achieve this complete hardware visibility, peripheral functionality can be captured in a graphical finite-state machine (FSM) language that combines graphical flowchart constructs with the execution power of ANSI C. This type of graphical language simplifies descriptions of concurrent hardware interactions and dramatically increases the MC visibility and controllability.

Technorati Tags: distributed computing, multicore

Powered by ScribeFire.

The future of multi-core by CPU designers

The EE Times online has an interesting review of the International Solid State Circuits conference in San Jose (Feb 5). Senior staff at AMD, Renesas, IBM and Tilera shared their view on the design of the next multi-core systems. Here are some interesting extracts about the role of software following multi-core architecture developments:

Shekhar Borkar, director of Intel Corp.'s Microprocessor Technology Lab, said microprocessor cores will get increasingly simple, but software needs to evolve more quickly than in the past to catch up.

"A core will look like a NAND gate in the future. You won't want to mess with it," said Borkar. "As for software, the time to market has been long in the past, but we can't afford to let that be the case in the future," he added.

Dave Ditzel, former CPU architect at Sun and founder of Transmeta, agreed. A member of the audience, Ditzel told the panel he helped design Sun's first 64-bit CPU then waited nearly ten years before commercial 64-bit operating systems became available.

Ditzel borrowed a metaphor from Berkeley computer science professor Dave Patterson to describe the current situation.

"With multi-core it's like we are throwing this Hail Mary pass down the field and now we have to run down there as fast as we can to see if we can catch it," Ditzel said.

Technorati Tags: distributed systems

Powered by ScribeFire.

What makes a successful entrepreneur?

A Marginal Revolution post linking to a paper from the World Bank analyzing Brazilian entrepreneurs. Cowen highlights some observations in his post but what I found most interesting (and controversial) is the following:

Interestingly, we find that having family and relatives run a business not only increases the probability of a respondent to be an active entrepreneur but also (and to a significantly larger extent) the probability to be a “failed entrepreneur.”

The most striking differences between active and failed entrepreneurs are as follows. Failed entrepreneurs are significantly less risk-taking. Interestingly, they report to have been significantly more often among top 10% in school even though they exhibited the lowest cognitive test scores. The low actual test scores point to the likely overestimation of failed entrepreneurs’ self-reported performance in school.

Overall, the results suggest that social networks play a big role in the decision to become an entrepreneur but not in determining whether entrepreneur will be successful. In contrast, the absence of risk-taking and greed, poor cognitive abilities, and over-evaluation of one’s self seem to be the main reasons to quit entrepreneurship. These results are similar to what we have found in Chinese survey (the data from Russia are not comparable).


Technorati Tags: entrepreneurship

Powered by ScribeFire.

80% of commercial apps to use open source by 2012

A few days ago I posted about open source software gaining ground in mayor IB projects. Today I came across a post on arstechnica.com that reviews a research by Gartner claiming 80% of commercial apps will use open source by 2012. I have a better time estimate for this process to happen: 2008. Open source tools are so widespread in software companies that I even wonder who is developing without using any (my thoughts run to some obvious hits, like Linux and g++, to more infrastructures-oriented products such as Samba and MySQL).

The Gartner gurus comment:

By 2012, 80 per cent of all commercial software will include elements of open-source technology. Many open-source technologies are mature, stable and well supported. They provide significant opportunities for vendors and users to lower their total cost of ownership and increase returns on investment. Ignoring this will put companies at a serious competitive disadvantage. Embedded open source strategies will become the minimal level of investment that most large software vendors will find necessary to maintain competitive advantages during the next five years.

Thanks Gartner. The report goes on with a number of IT forecast which are quite fun to read. For example, they have a take on Apple doubling PC market share by 2011 and the number of 3-D printers growing 100-fold through 2011. What I found more interesting is their view about software distribution, which should move from a license-based model to a service subscription charge (SaaS), an approach apparently endorsed by Google and Amazon:

By 2012, at least one-third of business application software spending will be as service subscription instead of as product license. With software as service (SaaS), the user organisation pays for software services in proportion to use. This is fundamentally different from the fixed-price perpetual license of the traditional on-premises technology.

Link to the original Gartner post.

Powered by ScribeFire.

Technorati Tags: open source

Simple equity market neutral fund

Mr. Hui out at the HSUTM blog has an interesting post on a simple market neutral portfolio:

1. Buy the top large cap Growth and Value equity funds, as ranked by Morningstar
2. The funds must be no-load mutual funds, have assets of at least a billion dollars and expense ratios less than 1%
3. Short the S&P 500 Spyder (SPY) against the portfolio
4. Re-balance the dollar amounts allocated to the funds monthly and re-balance the fund components annually

For the period from December 1998 to Janaury 2008 the synthetic equity
market neutral portfolio showed a very respectable annualized return of
6.4% (after fees) and a Sharpe ratio of 0.9.

Technorati Tags: investing, market neutral, HFRX

Powered by ScribeFire.