melonakos

It's not "anti-intellectual", it's a refreshing dose of anti-snotty-ism (not typically synonymous with cities containing top research universities).

Yeah, I started by saying shoot an email, but realized I didn't want to post my email address here :)

Xcelerate, shoot me a tweet to talk startups (@melonakos)

Costs are low enough in Atlanta that you can actually start your business without raising capital. Unthinkable, at least for an extended period of time, in most other cities.

Most businesses shouldn't raise money either because they're not going to grow fast enough to fit the VC model or they're going to dilute the founders so much that the wind gets sucked out of the sails. For example, learn from an Atlantan that just sold his Pardot no-investor-raised-money business for $95 million, http://davidcummings.org/2010/05/08/when-raising-money-makes....

Every city has pockets that are bad and pockets that are safe. Atlanta has tons of safe places.

Lame class warfare article. The left is full of whining idiots like this.

Higher preschool enrollment will not help. Fewer split homes and dual-income families with kids is the only way to ensure kids get the nurturing they deserve.

How is it possible to have negative points?

The jist of this article: "Hey, look at me! I'm cool because I am not patriotic, nor do I respect several of the rockbeds of American culture. I'm multinational."

As real Americans would say, "Whoop-de-doo!"

This is a follow-up post related to an NVIDIA blog post from one week ago: http://news.ycombinator.com/item?id=3796039

Aliens Cause Global Warming: A Caltech Lecture by Michael Crichton, http://bit.ly/vZX5si

Not all scientific fields are created equally. Some simply have access to better experimentation methods than others. Climate scientists have ridiculously poor measurement methods, historic records, and computational models to possibly be coming up with any conclusions.

An article on this topic: Err on the Side of Bologna – According to Scientists, http://bit.ly/wYbEIB

What world have you been living in that you think climate science is as "settled" as the speed of light limits?

Over the last 5 years, I've seen a ton of hot air blown about wrt to auto-GPU code generation. The latest hot air is about how magical directives make everything run fast.

Truth is, compilers and code generators are crappy.

If you really want to get good performance, you either have to write your own low-level GPU kernels, or use a library of functions that have already been written at a low-level.

All other hot air, while interesting, has yet to be proven at scale on more than a few limited use cases.

Another disclaimer: I work on this, http://accelereyes.com/arrayfire

right :) i fixed it.

Also, OpenCL is not going away, even if someone figured out how to get CUDA code to run well on ATI GPUs. OpenCL is gaining a lot of traction by mobile GPU vendors too (e.g. ARM Mali, Imagination PowerVR, Qualcomm Adreno, etc)

IMO, open sourcing their GPU libraries would be a much bigger deal than only open sourcing the compiler. I would like to see CUBLAS, CUFFT, CUSPARSE, CURAND, etc all get opened up to the community.

The pain is not in compiling GPU code; rather, the pain is in writing good GPU code. The major difference between NVIDIA and AMD (and the major edge NVIDIA has over AMD) is not as much the compiler as it is the libraries.

Of course, I'm biased, because I work at AccelerEyes and we do GPU consulting with our freely available, but not open source, ArrayFire GPU library, which has both CUDA and OpenCL versions.

If in MATLAB, use Jacket. It's better than what MathWorks offers, http://accelereyes.com/compare

If in C, C++, Fortran, Python, use LibJacket. Learn more, http://accelereyes.com/libjacket_tour

Disclaimer: yes, i work on both of those :)

I'm definitely partial :) I was relying on my profile to make it clear, but that's a bit buried.

BTW, you're profile has me cracking up in how you share your email. I tweeted it here: http://twitter.com/melonakos/status/11101399047737344

Do you think GPUs in R would be useful? We're heading that way with LIBJACKET.

The MATLAB GPU implementation is seriously flawed. For a better approach, see http://www.accelereyes.com/products/compare

No irony here. Just proud of the great American experiment and wish more people would embrace those virtues which lead to great outcomes.

Dependence on big government saps those virtues from those governed.

I did a very similar blog post back in August and took the other side, here - http://www.melonakos.com/2010/08/02/wikileaks/

I think there is a law of diminishing returns wrt working long hours. Is working 18 hr days really more productive that 14 hr days? That train of thought is what makes it not so insane.

I totally agree and wrote a similar piece a few months ago, calling it "Wallowing in Education", http://www.melonakos.com/2010/07/25/wallowing_in_education/

You can use Jacket for GPU matrix multiply in MATLAB and a host of other GPU accelerated functions (see http://www.accelereyes.com).