dhruvbird

I forgot top update this thread. I've taken my GRE subject test in CS and got a score of 850 (94%ile). I hope this adds value. Only time will tell I guess.

Hello Siddharth, Great score on the GRE!! I've just completed applying and I don't think all my scores have reached the universities. I'm not sure how many universities that I have applied to give out rolling admissions. Most of the places say that the admission decisions would be made around March/April. I too am applying first time, so I don't have much of an idea about these things :-) Regards, -Dhruv. update: I haven't applied to madisson, unc, ncsu, ohio SU, vtech, ufl, gatech or umich.

@CalmLogic & ic10503: Thanks for the wishes :-)

Hello, I just got my scores from the october exam. Score:850 %ile below: 94 total attempted: 65 correct: 60 incorrect: 5 raw score: 59 Is this score decent to apply in any of the top univs. or should it have been > 95%ile??

Does anyone know if scores are available on the phone now?

All the best to everyone!! Do well!!

Thanks! It seems that ETS is showing very old stats. Since it includes stats from 2004, there is a very high chance of dilution and there may in fact be a steady drop of people taking the test in CS on a year-on-year basis. We'll never know unless ETS releases some stats.

Yes, that's true :D I took the exam in Nepal, and there were just 3 people (including me) taking the CS test and 15 overall (taking any sort of subject GRE test). Well, then it seems that with the same score, you stand a greater chance of getting a higher %ile if you take the test in November as opposed to October. Does ETS release any other stats (like performance of a country?)

If ETS doesn't report how many people took the test then how did some people on this forum arrive at the conclusion that much fewer(100) people take the test in October as opposed to November(500)? What are the data points?

I am guessing this is just another way to make more money. I mean how much time does it take to have scores available online? Besides waiting is probably not an option for me since mailing anything to India takes at least 3 week and they won't use speed-post.

Actually, I'm a tad bit worried because I've been reading that due to the low participation on October (as opposed to November), it is very hard to get a good %ile. I can imagine if say there are only 50 people taking the test, every lower grade will result in a 2 %ile drop, which means that only 5 people will get 90%ile and above!! Does ETS do anything to mitigate this?

By mail do you mean email or snail mail? Also, if they are willing to give the score over phone, will they update it immediately online as well? Please do let us know how much you scored :)

q1. When will the scores for the october exam be available? q2. I was reading this thread and it seems that there is some variation between raw score to the %ile mapping on a year-on-year basis. http://www.www.urch.com/forums/gre-computer-science/78897-post-agre.html Please could some of the regular test takes try to give a rough estimate of what %ile to expect with a raw score of: 30 35 40 45 50 55 60 (added later) q3. Does ETS report how many people took the test in october? Regards, -Dhruv.

Count me in!! ;)

Hello Focus, Isn't it frowned upon to do so? I mean lots of these university sites (and also come professor's home pages) ask you to explicity NOT do so. If someone is interested in helping out, I am guessing they would help us in the forums such as this and edulix.

a gentle bounce. Can anyone suggest? CalmLogic?

Hello, I'm applying to MS in CS and PhD in CS for Fall 2011. Some information: GRE: 1430 - Q:800, V:630, AWA: 4.5 Toefl: 112 (with 23 in speaking) Awaiting GRECS subject test scores. I have 4 years of work experience and have taught courses on OS and Systems programming for 8 months. My projects include: 1. A distributed database: TDDB Home Page 2. A peer-2-peer file system: http://http://p2p-fs.sourceforge.net/ 3. An auto-complete service for find-as-you-type kind of systems 4. A distributed message queue: pymq - Project Hosting on Google Code I want to apply to 3 out of the 5 places mentioned below for a PhD in databases and distributed systems at these places after looking at the research happening there. 1. Yale -- sure 2. Berkley 3. MIT 4. Gatech 5. Umich Ann Arbor 6. CMU I also want to apply for an MS to about 10 out of the universities listed below considering that it would be very hard to directly get an admit into one of these PhD programmes: 1. utah 2. vtech 3. SUNY SB 4. umd (umcp) 5. ufl 6. univ. of arizona 7. ut austin 8. umn 9. winsonsin (madison) 10. ohio state univ. 11. ncsu 12. unc - Chapel Hill 13. indiana bloomington So basically, I am looking to do a good MS degree for cheap and get some good research and papers published so that I can then apply for a PhD after that. My fields of interest include databases, distributed systems, information retrieval and data mining. Please could you give me some suggestions about university applications based on the information above. Also, if you have any other universities that you think I should check out, please do let me know about them. Thanks! -Dhruv.

While I was studying for the test, I found it a bit quirky to note that DCFLs (Deterministic context-free languages) are NOT closed under concatenation. You can find the table of closure properties of various languages here: Formal language - Wikipedia, the free encyclopedia or use these notes that I made up: closure_and_set_properties_of_languages.odt At first sight, I though that DCFLs should be closed under concatenation, but it is not so. However, it is possible that the suffix of a former DCLF matches the prefix of a later DCFL. The DPDA would not know which part it is currently processing and would need non-determinism to correctly recognize the strings of the newly formed language. Hence, DCFLs are NOT closed under concatenation.

I've tried putting together a wiki page for people who want to get more familiar with the test here: http://grecs.wikispaces.com/

Another thing to note would be that according to this wikipeia article: Euclidean algorithm - Wikipedia, the free encyclopedia , the Euclidean algorithm always needs less than O(h) divisions, where h is the number of digits in the smaller number b. Meaning that if Big-Oh is abused to mean a tight bound (as it generally is), then the complexity should be O(min(log(lowValue, highValue)) and not O(max(log(lowValue, highValue)). However, technically speaking, O(max(log(lowValue, highValue)) is perfect.

Hello Greg, Even I would interpret it the way you did. Maybe Chris (the author of Titanium Bits) meant to write something else, or we are both not getting what he has to say. The complexity of Euclid's algorithm is particularly hard to analyze though. I have yet to work it out.

[A] Technique for using the Master Theorem where it isn't immediately apparent. Let lg(x) = log2(x) Let's look at a simple recurrence equation where we think that the Master Theorem may not be applicable: eg 1. T(n) = T(n-1) + n Let n = lg(k) Therefore, T(lg(k)) = T(lg(k/2)) + lg(k) Let T(lg(k)) = F(k) Therefore, F(k) = F(k/2) + lg(k) Using the Master Theorem, we get: O(lg2(k)) Reverse-substituting (k = 2n): = O(lg2(2n)) = O(lg(2n))2 = O(n2) eg 2. T(n) = 2T(n-2) + 1 Let n = lg(k) Therefore, T(lg(k)) = T(lg(k/4)) + 1 Let T(lg(k)) = F(k) Therefore, F(k) = 2F(k/4) + 1 Using the Master Theorem, we get: O(k1/2) Reverse-substituting (k = 2n): = O((2n)1/2) = O(2n/2) = O(2n) Solving seemingly complicated recurrence equations eg 1. T(n) = T(n1/2) + 1 Let n = 2k Therefore, T(2k) = T(2k/2) + 1 Let T(2k) = F(k) Therefore, F(k) = F(k/2) + 1 Using the Master Theorem, we get: O(lg(k)) Reverse-substituting (k = lg(n)): = O(lg(lg(n)) eg 2. T(n) = T(n/2) + T(n/4) + 1 Let n = 2k Therefore, T(2k) = T(2k-1) + T(2k-2) + 1 Let T(2k) = F(k) Therefore, F(k) = F(k-1) + F(k-2) + 1 We know that the recurrence above is that of the Fibonacci sequence. Hence is equal to: O(2k) Reverse-substituting (k = lg(n)): = O(n)

Hello, Does anyone know if it is correct to say that all languages in NP are Recursive Languages? I would expect them all to be decidable by a TM, but do any counter-examples exist?

Hello, How many people took the test this month (October) and how many are planning to take it in November? I am just back from taking the test.