Forgot your password?
typodupeerror

Comment: "High level" programming environment? Sigh. (Score 2) 76

by nxmehta (#36233796) Attached to: Cray Unveils Its First GPU Supercomputer
The fact that writing C and Fortran code using a message passing library constitutes a high level programming environment is a complete indictment of the sad state of parallel programming today. Seriously, do you want to be programming complex parallel algorithms on HPC machines using Soviet Era technology? I've tried that and it made me want to jump out a window. It's about as easy to program in this type of an environment as it is to program an FPGA (hint: it's a pain in the ass).

Techdirt: This Is The New EMI? Trying To Personally Bankrupt CEOs Of Companies It Doesn't ->

From feed by techdirtfeed
Last year, a private equity firm took over the major record label EMI and announced that things were going to be different. It wasn't going to be anti-fan. It was going to look on the success stories of Radiohead and Nine Inch Nails not as a threat, but as an opportunity. It even threatened to leave the RIAA and the IFPI unless it pulled back on suing fans. It then went out and hired internet savvy executives who would (hopefully) contribute a different perspective to the running of a major record label.

But, in the midst of all of this, it hasn't really backed down from a variety of ridiculous lawsuits. For example, it's been a part of a lawsuit against an ISP that refuses to spy on its users and cut off those who do unauthorized file sharing. We also noted last year that it had sued Michael Robertson's startup, MP3Tunes. Now we tend to have a lot of fun accusing Robertson of having the same marketing strategy with every company he starts: piss off some established company, get sued. It seems to happen with pretty much every company he starts from MP3.com to Lindows/Linspire to SIPphone. So it really wasn't a huge surprise to see MP3Tunes sued -- though, it's difficult to see how a personal music storage locker that doesn't allow sharing could possibly be infringing.

However, now it appears that the "new" more "friendly" EMI isn't just suing MP3Tunes. According to Boing Boing and Michael Robertson, it's trying to extend the lawsuit to go after Robertson personally, saying that he should be personally liable. As you probably know, one of the purposes of a corporate structure is to limit the liability of the executives of a company. To go after Robertson personally makes very little sense unless the idea is to intimidate. Many executives will quickly settle in such circumstances so as to not open themselves up to such a huge liability. On top of that, the chilling effects are tremendous. Others won't even think of starting innovative services, for fear of being personally liable in a lawsuit.

Unless EMI pulls back its lawyers, I think we can safely conclude that the "new" EMI hasn't really changed much from the old EMI.

Permalink | Comments | Email This Story


Link to Original Source

Comment: Ah, but that's the point! (Score 1) 154

by nxmehta (#23285086) Attached to: NVIDIA Shaking Up the Parallel Programming World
The entire reason why CUDA works and is powerful is exactly because it is limited. Nvidia knows that there is no silver bullet. They're not claiming that this is one (David Kirk has said so himself at conferences). CUDA is a fairly elegant way of mapping embarrassingly data parallel programs to a large array of single precision FP units. If your problem fits into the model, the performance you get via CUDA will smoke just about anything else (except maybe an FPGA in some scenarios).

Your notion about particular models making some parts of parallel programming easy while other parts are hard is what people really need to learn to accept about parallel programming. If you're expecting a single model to make everything easy for you, trust me, stop programming right now.

You need to pick the programming model that matches the parallelism in your application- there will never be one solution. When sitting down to write code, you have to ask yourself: what is the right model for this algorithm? Is it:

Data parallel (SIMD, Vector)
Message Passing
Actors
Dataflow
Transactional
Streaming (pipe and filter)
Sparse Graph
Etc...

There are many models out there, and many languages + hardware substrates for these models that will give you orders of magnitude speedup for parallel programs. They key is to just to sit down, think about the problem, and pick the right one (or combinations).

The real research focus in parallel programming should be to make a taxonomy of models and start coming up with a unified infrastructure to support intelligent selection of models, mixing and matching, and compilation.

Badges? We don't need no stinking badges.

Working...