I guess that's the tradeoff for the so-called "free" open source software. You remove the burden of paying costly licenses, but then you're obliged to invest in the training/people/resources to learn the environment, from the inside/out, and to contribute back to the community. The upside is there are many great resources and experts to tap in the community to learn the technology and get an edge on extending it. But again, that takes time, and in the engineering world, time is money.
I think people ONLY use OpenFOAM because of the cost. Single computer codes like Fluent cost $25k plus per year. CD- ADAPCO one of the leaders in automobile computing charges $10k per 100 hours on a multiprocessor machine.
Jim: Great perspective. I'm curious about the price issue you raise, though. From everything I've ever heard about open source software, the fact that's is "free" is somewhat of a misnomer. Most people say there are hidden costs around getting the support and resources, including folks who are comfortable getting their hands dirty with code. On the other hand, I guess people diving into CFD are the types that don't mind--more likely, relish--getting their hands dirty with code. Beyond cost, though, what other features/capabilties make OpenFOAM stand apart from the rest of the CFD pack?
No doubt in my mind whatsoever that open source is the next big thing in CFD. We have used virtually every major brand of software: Fluent, Fidap, Nekton, Rampant, Adina, Cosmos, and so forth; yet we've almost exclusivley migrated to OpenFOAM. For so many reasons. Not the least of which being the price is hard to beat! But most importantly is the flexibility, and freedom to get your "hooks" into the code. Congratulations to SGI, you've bagged a winner!
Ivan, I assume you're talking about supporting GPUs like those from NVIDIA. While they didn't mention that in the FAQ or in subsequent interviews, a quick Google shows that there seems to be available plug-ins (remember, we're talking open source) that "support the execution of simulation on multi-gpu systems." I would think that and OpenFOAM's parallel processing capabilities could address some of those traditional CFD performance issues.
I used SGI workstations some time back for doing 3D modeling for USN ship design at Newport News Shipbuilding. Their machines were impressive but the rise of commodity hardware CISC systems seems to have made the RISC based SGI machines less attractive.
The availability of high end graphics cards that can perform specialized processing that was originally developed for games to be utilized in other ways is a recent trend. Several of the distributed computing projects like SETI, the LHC and Einstein can make use of the graphics cards. I am curious if the Open source CFD makes use of any of these high end cards. I have 5 that are dedicated to specialized operations that will eventually be turned over to this kind of distributed processing to support some of the projects within BOINC.
Modern multicore cpu's are entirely outclassed in some compute intensive operations such as CFD by these new graphics cards that feature hundreds if not thousands of stream processors. The processors perform simple operations millions of times in parallel.
SGI will do well integrating this kind of hardware and software into a complete dedicated system. Such tools will greatly reduce the effort required to perform the necessary computations effectively and efficiently. The efficient part is significant since the proper utilization and tuning of the graphics cards can reduce the power inputs. I have the highest electric bill ever last month as a result of the power usage and cooling required for the computers.
Robots that walk have come a long way from simple barebones walking machines or pairs of legs without an upper body and head. Much of the research these days focuses on making more humanoid robots. But they are not all created equal.
The IEEE Computer Society has named the top 10 trends for 2014. You can expect the convergence of cloud computing and mobile devices, advances in health care data and devices, as well as privacy issues in social media to make the headlines. And 3D printing came out of nowhere to make a big splash.
For industrial control applications, or even a simple assembly line, that machine can go almost 24/7 without a break. But what happens when the task is a little more complex? That’s where the “smart” machine would come in. The smart machine is one that has some simple (or complex in some cases) processing capability to be able to adapt to changing conditions. Such machines are suited for a host of applications, including automotive, aerospace, defense, medical, computers and electronics, telecommunications, consumer goods, and so on. This discussion will examine what’s possible with smart machines, and what tradeoffs need to be made to implement such a solution.