How difficult is it to get into a good, computer science graduate program with an undergraduate GPA below 3.0?

If this gives you any hope, I might mention that (a CS professor at Utah) explicitly says that he doesn't care about grades. See http://matt.might.net/articles/how-to-apply-and-get-in-to-graduate-school-in-science-mathematics-engineering-or-computer-science/ What doesn't matter GPA? I don't care if it's 2.0 or 4.0. I won't even look at it. The school you went to? I'll judge you the same whether you went to Nowhere State U or a top-ten school. Transcripts? Never seen one. GREs? Irrelevant. Where you work/worked? Unless it's a research lab, it's not important. I don't think these items have much predictive capacity as to whether or not someone can complete a Ph.D. A word on GPA cut-offs I discovered through feedback that some schools (including Utah) have a GPA cut-off. I think GPA cut-offs are absurd. Of course, GPA cut-offs are not hard. In practice, there is a way to override them, but it probably requires a professor going to bat for you and getting the right bits flipped in the university bureaucracy. So, if you have a low GPA, mention it after you've piqued a professor's interest, and ask if you think it will be a problem during the admissions process. [One of Utah's Turing award winners was a "special-case" admit for grad school because his GPA was below the cut-off. We're glad we didn't reject him.] If I really like you, I'll go to bat for you. Now, Utah certainly isn't a top-5 CS graduate program, and 's philosophy is probably different from that of most. But I've still posted this just to let you know that there are professors who really do exist who don't care much about undergrad GPA. In summary, the other answers are most likely right. Also - there's someone from Stack Exchange (Jeff Erickson, as Mirko mentioned in this thread) who got into UIUC CS with a < 3.0 GPA.

Several years ago Jeff Erickson wrote a good answer to this question in his blog.  Jeff is now a Theory Professor at UIUC, but in undergrad had a GPA of 2.4/4.0. http://3dpancakes.typepad.com/ernie/2005/03/re_phd_with_low.html He tells his story and I think it provides several suggestions on how to deal with the problem.

It is very challenging and will require offscale recommendations. Some schools ask the recommender to classify you as a student: are you the best student they encountered this year, the best student in the last decade, or the best student they have ever encountered in their entire career as a professor? If you think the recommenders will check the last box, and they are people the admissions committee can trust, you may have a shot. If you have some special, distinguishing accomplishments under your belt that suggest you are poised to change the world, you may have a shot. Don't be afraid to ask your recommenders frankly whether they can give you a strong recommendation; if the answer is no, they may be relieved that you were forthright. I am not sure that job experience at Google/Facebook/etc. will make much difference; the committee is looking for indications that you are a special student in the top 4% or whatever share they admit. If you had lower than a B average as an undergrad, you will need considerable "special spark" in your portfolio to overcome that handicap. I also would urge you to think about your goals in postgraduate education and where you might be able to achieve them. If you do want to pursue doctoral studies in computer science, there are plenty of places that will provide an outstanding opportunity to do so that are not Cal, MIT or Stanford. It is certainly possible to go to grad school in a place that is not in the absolute top five, do well and distinguish yourself (or get lucky and stumble on a hot area), have a great dissertation, and then have an amazing career. The three schools you named are certainly good and can afford to be the choosiest, but they are not the only game in town.

I just got into NYU with seriously terrible GRE scores and a 3.0 undergrad GPA. I realize the program is different, but anything can be helped by a well written essay. I got into every school to which I applied by this alone. My recommendations were good, but they highlighted my professional strengths and not my academic achievements. Here is my source material: http://www.amazon.com/Graduate-Admissions-Essays-Fourth-Edition/dp/1607743213/ref=sr_1_2?ie=UTF8&qid=1332737093&sr=8-2

If by "good" you mean a place like Stanford or Berkeley, the answer is almost completely impossible.  Yeah, if you had great GRE scores, some first-author publications in relevant journals, and glowing recommendations there's a small chance you could get accepted, but if you were smart enough to be better than everyone else in those regards you'd probably have a higher GPA.  Also, no matter how high your test scores or how great your recommendation letters, you're competing against people all over the world.  It's likely that spaces in the top graduate programs will be filled by people who have test scores and publications that are better than you can ever imagine yourself having in addition to near-perfect GPAs.

I'd say virtually impossible. Undergraduate scores are an important predictor of success in graduate school. Spectacular recommendations from CS professors at a your elite undergraduate school might help, but with a 3.0 they aren't very likely (unless you had a 4.0 in CS and a 2.5 in everything else). I have advised a student with a double major in math and computer science, a 4.0 in mathematics a 3.67 in CS and overall at a top school --  and extra credentials in math and as a teaching assistant. He did not get admitted to the top tier in CS (he did get admitted with financial support from more than one of the top 25, however). Job experience is unlikely to affect admissions unless you have been engaged in breakthrough research in the area you wish to pursue.

Don't limit your options with top schools. There are also good schools that aren't Ivy League. See this list of the http://bitly.com/1tAKBG1. But do try anyway by compensating what you lack with other credentials like skills, work experience, and most specially recommendations from people in the academe.

Many students start their college research with rankings. That’s all well and good. You’ve got to start somewhere. But overall there is no ranking system, no acceptance criteria, no matter how perfect, is going to be able to tell you what university is best for your future.  I recommend you to read this link in detail. This will answer some part of your question.