manvsmachine

Very true. It is for the same reason that the article shows GPU progress outpacing CPU's in regards to FLOPS. The limited scope of a GPU's functionality allows most of the processor's silicon to be arithmetic units. CPU's, on the other hand, have huge sections of their real estate dedicated to handling conditional logic (something most GPU's fail miserably at), legacy instruction handling, virtualization, etc.

Not anymore, I'd assume. It's running 2 year old hardware by now and Fermi-based GPU's pretty much trounce the GT200 family when it comes to double precision. That said, another good example of how GPU computing can provide a lot of horsepower for (relatively) little cash, especially for certain classes of problems. I built a Tesla Desktop SuperComputer (Dual Xeons, 3 Tesla's and a Quadro 4800) for my school's Computer Vision lab back in '09, and it cost us ~8K for a ridiculous amount of performance.

Just a thought: maybe that's a feature, not a bug. Seems like a quick and dirty way to put up a site that people couldn't get to unless they're already know what they're looking for (which would imply that you want them to find it). Probably not the intent, but I'm kind of intrigued by the idea of it essentially being an Internet stash house.

Here's what I used while doing an undergrad independent study a couple years ago (except for the books; at the time, books on the subject didn't exist yet):

Online Tutorials:

Dr. Dobbs - CUDA, Supercomputing For the Masses: http://www.drdobbs.com/high-performance-computing/207200659

GTC Video Tutorials / Presentations: http://www.nvidia.com/object/gpu_technology_conference.html

University Courses:

GaTech ECE4893 - Multicore and GPU Programming for Video Games (not CUDA specific, but covered): http://users.ece.gatech.edu/lanterma/mpg/

UIUC ECE498AL - Applied Parallel Programming: http://courses.engr.illinois.edu/ece498/al/

Books:

Programming Massively Parallel Processors: http://www.amazon.com/Programming-Massively-Parallel-Process...

CUDA By Example: http://www.amazon.com/CUDA-Example-Introduction-General-Purp...

If you decide to go the CUDA route, as opposed to OpenCL, then NVIDIA has a lot of solid documentation that will help you get to a "Hello, World" level pretty quickly. This StackOverflow thread should help if you choose to go with an AMD OpenCL solution: http://stackoverflow.com/questions/780978/opencl-books-tutor.... Personally, I would stick with NVIDIA; they seem much more dedicated to advancing the cause and you have support for both API's.

If you're still talking about about NVIDIA cards, then no; their OpenCL implementation is based upon the CUDA architecture (technically, CUDA is the device architecture that supports GPGPU programming, not the language itself). That said, videocards are cheap these days, as long as you don't need the high-end: you can snag an entry-level Fermi card (GTS450) for right around $100 these days, or less if you go with something GT200-based.

courtesy of tptacek's HN profile: http://amzn.to/cthr46

Really good book. GA Tech used this as their "Intro to CS" text until they decided that learning scheme was too much for business and inta majors (everyone in the school had to take the class). I'd recommend this as a potential alternative for those thinking about learning to program using SICP. The latter dives into some more advanced topics, but this is much more accessible to newbies imo.

From my (admittedly limited) experience, yes, you will want to eventually get a Mac of our own for development. I'm not sure if there is a way to create .dmg packages otherwise, but you will definitely want to be able to take advantage of the Cocoa libraries if you're doing GUI apps. The most common recommendation I hear for strictly cross-platform development is to get a Mac Mini; it can handle pretty much any "normal" development tasks you'd need it for.

And, apparently, Linux users are not. I find that interesting, considering that one of the most prevalent arguments for not offering software for Linux is that its users refuse to pay for anything.

This has been done to death, do a site search and you'll find tons of suggestions.

That said, http://khanacademy.org

I agree with the overall sentiment, but the Nexus One doesn't really fall into the latter category considering that they just released the Nexus S. They never abandoned the Nexus line itself, just their online-only distribution method.

Because if you don't, you'll be dead within the next day: http://godheadv.blogspot.com/2010/04/abandoned-on-everest.ht...

Your point still stands though.

A commonly reiterated finding is that, while having more money can indeed contribute towards a happier life, it only does so to a point. According the the article, Zuckerberg's speculatively worth roughly $6B; if this proves accurate and he gives 90% of that away, he's still sitting on a small fortune. I'd hope that he is capable of figuring out how to make do with "only" $600M or so.

As someone who always had that problem, I think it's because I always mentally picture a command as saying "do <action> from <arg0> to <arg1>", ie, "copy this file to that file". But this construct doesn't hold up for linking, so I just have to remember it arbitrarily by remembering cp. function(src, dest) just generally seems to be the unofficial "right way" of ordering things.

you may want to check this out: http://www.aaronboodman.com/2010/10/wherein-i-help-you-get-g...

It seems that your first step is to to find good C++ resources: http://stackoverflow.com/questions/155762/best-c-resource

Hope this helps.

At first, I was going to comment on the blatant linkbait approach of this article, but then I realized that there actually is a lesson, albeit a simple one, behind the gimmick - go out and find what you want. Don't wait for it to find you, cause there's a good chance that it never will. Want to marry a beautiful Brazilian women? Suck it up and move to Brazil; now half of it is already taken care of. Is your startup's location hurting its chances of getting traction? Take it to SV / Boston / DC / wherever is applicable for your situation. Your customer base doesn't know about your product? Then find them and let them know.

We often start to think that success is simply the result of hard work or building something awesome, but we overlook how important it is to maneuver ourselves into a situation where we can capitalize on all that work and preparation. There is no substitute for good old-fashioned legwork.

I didn't even know about the challenge until today, but trying it just now made me realize that Chromium's built in dev tools are lacking some must-have functionality.

The form was broken for me; it wouldn't accept a graduation year of 2010.

Apple doing it and Spotify doing it are two completely different things, traffic-wise. Assuming that Apple is basing their streaming solution off of Lala's model (not a given, but why else acquire them?), we're talking about giving every iTunes user streaming access to their entire music library, including songs that weren't purchased through iTunes (which would require a serious amount of cloud storage). Spotify reached one million users sometime last year; there are how many millions of iPhones and iPod Touch's out in the world?

I saw the same . I understand his attitude somewhat - imo, too much of the economy is focused on marketing unnecessary products to people that barely want them. That said, making and moving things is a critical element societal of advancement. The entire reason people can read his blog post is because some companies decided to make some things. I consider myself an academic at the core as well, but the world runs on more than just ideas.

The line that stood out to me most:

The amount of difference between a good programmer and a great programmer is going to get wider, in terms of the amount of value that you can create.

In the case of the Fermi-class cards, you're paying for more memory (3 or 6GB vs 1 or 2), stability, tweaked drivers, and (possibly most importantly) error correction. It's really the same argument as for workstation graphics cards - consumer class cards are designed to sacrifice floating-point accuracy for maximum speed, and in some cases, this is unacceptable.

That said the 20- series is really freaking expensive. When I got my C1060's, they were ~$1500/per retail; I think the C2070's are nearly twice that.

Video games. Seriously. Here's why:

1) Hardware knowledge. I didn't start programming until I started college, but I quickly realized that, growing up as a pretty avid gamer, I still had one advantage that a lot of the CS majors didn't have: I knew more about computers than they did. Games are some of the most demanding applications being created today, and, as a result, I had spent a lot of time reading hardware reviews and benchmarks. This resulted in my having a somewhat innate understanding of how data is passed around inside a computer and how this affects performance. It takes time to understand on a deep level how your code is interacting with the hardware, but when you see the frame rate tank because you used an inefficient algorithm somewhere, it sticks with you instantly.[1]

2) Instant gratification. Much like web programming, game programming allows you to see the result of your work, immediately and incrementally. And it looks good, because you get to use professionally developed resources (models, textures, etc). A side benefit is having good looking programs to put in your portfolio.

3) Wide variety of problem domains. Once you get familiar with programming in general, you can then work on pretty much anything you want: 3D graphics, physics simulation, AI, networking, UI / UX , HCI, etc.

4) It's fun.

Try giving Unity a shot and see how you like it: http://unity3d.com/unity/download/ . It's multiplatform and has API's in C#, JavaScript, and Python, so you have an element of choice. If you're a Windows user, I know a few people who started out by playing around with XNA, which uses C#. Examples of free commercial game SDK's are Epic's Unreal Development Kit (http://www.udk.com/) and Valve's Source Engine SDK, which comes free with any Source Engine-based game.

[1] When working on a graphics project, I accidentally put the function that imports and initializes the texture resources into the draw loop, which meant that new memory was being allocated for them 60+ times per second. Needless to say, I became very aware of the problem when my fans started to spin up like jet engines.

Phoronix did seem to have some compelling evidence that Valve was at least toying with the idea. My guess is that Valve's statement doesn't mean that they aren't interested in the potential of Steam on Linux, but that they aren't actively developing for it and therefore do not want the public to have expectations for progress, release dates, etc. Companies do this all the time; one of the inside jokes of tech news seems to be that, if Apple or Sony take the time to officially state that they're not working on something, then they're probably working on it.

That actually makes a lot of sense. At the end of the day, RIM and MS have always shared the same target demographic for the most part - they are the de facto handset / OS of the enterprise. The iPhone has slowly been working its way into the enterprise sector, but they could easily lock down that entire segment by focusing on security + encryption and heavily integrating with Visual Studio / .NET. It would also allow Microsoft to benefit from having pseudo-vertical integration, as their best products seem to be the ones where they are developing for a single hardware platform (eg, Xbox, later Zune models, etc).

For some reason, that doesn't really surprise me. I mean, GMail was basically a whim, and I kinda got that same feeling when Wave came out. The Android acquisition could just be considered like a 20% project, but scaled to the CEO level.

I didn't get the idea that he was saying that 1200kcal was an everyday thing; if it was, then I totally agree. I can picture having 1200kcal days here and there when he's not really active though, like his aforementioned 12-15 hour hacking days. The main point being: listen to your body. I you feel like you really need to eat something, then eat something. If you don't, then don't.

Nice writeup Paul. One thing I'd point out is that your dietary needs will evolve as your body and metabolism does. Once your body adapts to its new state and you start gaining muscle mass, you'll actually start needing more calories and you can relax on the little thing like egg yolks, etc.

Another thing that you might want to consider if you're not the hardcore gym type is adding some less traditional forms of exercise. A lot of people really get into martial arts despite not liking to "workout"; I know you like driving, so I'd recommend seeing if there are any good karting options in SF. Karting is crazy amounts of fun, and its a great all-around workout: http://www.itv-f1.com/News_Article.aspx?id=46561&PO=4656...

I think it would work - essentially, this is the entire premise of the simulation market; the only difference is a lesser emphasis placed on motor skills due to it being an RPG. Mastering Gran Turismo / Forza requires players to learn quite a bit about real-world vehicle dynamics, for example. The caveat is that, in your example, the cumulative knowledge required to develop the game would be equal to mastery of every in-game class. You would need several teams of expert consultants to create and model scenarios, much like TV shows that deal with medicine, military, forensics, etc. Realistically, it would make a lot more sense to narrow the focus to one or a few domains.

In the spirit of a recently created post, how long did it take you to build this, from conception to release?