grabanca

Well, that's 13 not 30 :) Anyway I am thinking of the following arguments: 1. is it not normal for an infant to be distressed when hearing unknown voices? the argument does not show a comparison between these infants and other ones. 2. similar to above: what percentage of teenagers are shy overall according to researcher's standards. shy is not easy to define absolutely; it has to be defined relative to 'normal' kids. 3. the shyness of the teenagers might have a complete unrelated cause. those kids have something in common: they were born in fall of the same year and are probably from the same country so they might have been exposed to similar events in their childhood (an earthquake when they were young, parents stressed because of bad economy, different style of education) 4. even if children born during fall would end up being more shy than others, there might be other 1000 obscure reasons for that (like they were always first to have their birthday in the new school year).

Guys, thanx for participating in this great forum. I also got a 870 (97%) score. I didn't answer 5 questions but on the others I was pretty sure. I guess I may have made 1 or 2 mistakes... I had almost no clue when it came to network stuff (4 out of those questions) but I have the theory pretty solid. I think the trick is to not get into problems that are too difficult. Solve many simple problems and make sure that you can understand the problem, do the arithmetic and verify yourself in 2 to 3 minutes. Also, being able to understand someone else's code is crucial. Competing in TopCoder algorithm matches helped me practicing to understand ugly code fast. I studied for this for about a month and this forum was of great help. Especially the posts in 2003...

I just came back from the test site. Hope I did ok... hard to tell. Especially that it really matters what others did as well. Left 4 or 5 questions unanswered, all those TCP, Ethernet and stuff. How long until we find out the results?

1. I think the answer is i. Floating point numbers generally have a hard time doing addition of large numbers that ends up == 0. if sqrt(B*B - 4AC) / 2A is close to B, there may be a large relative error. 3. All such languages are regular. For multiple of 8 is obvious: all strings must end in 000. For the others, we may construct a DFA with at most 3, 5 or in general n states that records the remainder of the division in the number we read so far. Ex. n = 3 State | 1 | 0 | R0 R1 R0 R1 R0 R2 R2 R2 R1 Starting and accepting state is R0 (remainder 0) Ex. 1001 : R0->R1->R2->R1->R0 Accepted 1010 : R0->R1->R2->R2->R1 Not accepted (remainder 1)

I agree. I got it wrong first but it makes total sense. Thanks R0jkumar and Gttts23.

I would think that the answer is B) III only. But I am not sure as well. I think a crash means something like power failure. And the goal is not to have the information saved, which cannot be guaranteed, but to have the system in a valid state. The danger is that the power failure may stop the I/O in the middle. With atomic write, the I/O is either completed or not started which will leave the system in a consistent state. I am not sure how the other choices would help...

In a system using buffered I/O, which of the following are needed for ensuring a consistent state on recovery from a crash: I. Allocate more main memory for buffers II. Duplication of buffers in memory III. Atomic write of multiple buffers A. I B.III C. II, III D.I,II E. I, II, III

4. What is the min number of states required for converting ANY NFA with N states to a DFA: A. N B. N^2 C. 2^N D. 2N E. N! *

Anyone wants to have a shot at this? If L \subseteq {0,1}* is not recursive then which of the following are true: I. L' is not recursive II. L is contained in a recursive set III. L is infinite A. I B. III C. I,III D. I,II,III *

taro_curly and iamdaisy your expressions do not accept strings like baaaabaaaba. Here is my shot at it: (a + ba)*(b + eps) A "b" must be followed immediately by "a", unless it is the last letter. Does it look right to you guys?

My answer for Q2 is E as well, though I am not sure... Any number is ordered correctly relative to n-1 other items. It is also ordered correctly relative to 1/3 of the remaining items (because it is 3-ordered) which implies that a number is ordered correctly relative to |_4n/3_| items. Generally a number can be up to [2n/3] positions away from its place in the ordered array. For n = 1, 2n/3 = 1 as well. Ex. [2,1] is 2-ordered and 3-ordered, n=1 and the maximum distance is 1. I would choose 2n/3 if there were such a choice. Otherwise I think n makes most sense.

Consider the 2-ordered Array of size 10 (n = 5) 1 6 2 7 3 8 4 9 5 10 6 is put into position 2 instead of position 6 => distance is 4, which is greater than n/2. So the answer could only be n or 2n-1. Thinking about the fact that any number is sorted in raport to n other numbers leads to the conclusion that the answer is n.

An ambiguous grammar is a grammar that generates a language L and there exist one word w in L such that there are 2 or more different possible parsings for it. For example S-> SS | 0 is ambiguous. The string 000 is part of the generated language and can be parsed 2 different ways: 1. S->SS -> 0 S -> 0 S S-> 0 0 S -> 0 0 0 2. S->SS -> SS S -> 0 S S-> 0 0 S -> 0 0 0

Q1 needs 3 states. Do you have a solution with 2 states only?

Q2: State | a | b | q1 q2 qf q2 q3 qf q3 q3 q1 qf qf qf

L1 = (011 + 1)* L2 = 1* (0 + 1* + 0 + 1* + 0 + 1* + 0 + 1*)* 1* L3 = (0 + 1)* 1 (0+1)(0+1)

I think you should try. From what I read on some forums from past years, there are barely a few CS test takers every year in each test center. So I guess they will be glad to take your late registration fee for a free seat.

That makes sense. Thanks

I believe the answer to b) should be 200ns + 200ns * (1 - .75) = 250ns. Only the page references not found in TLB have to be read from the page table (50ns in average). Then, all memory references still have to be done to get the actual data.

Thanks CalmLogic, that helps a lot.

Hi, I did not see many of these type of problems but nevertheless... How much video memory is required for a 1280 x 800 resolution with 64 KB colors?

Consider a paging system with the page table stored in memory. a. If a memory ref. takes 200 ns how long does a page memory ref. take? b. If a TBL is used and 75% of all page-table references are found in the TBL, what is the effective mem. ref. time? (finding a page entry in TBL takes 0 time)

Hi guys, this is my first post. I have been reading a lot of questions from 2003 and this forum seemed to be much more active... Can we try to get it back to speed? Just 4 weeks left... I think the hardest part of the test is that there is a lot of breath to cover. If you understand the problem it is usually not too difficult. So I thought having a list of terms is important (saw these in problems a lot and sometimes I had to look them up). OS paging: page tables (inverted page table, multiple level page table), TBL, associative memory, LRU algorithm, FIFO, Optimal page replacement, Most Recently Used, Belady's anomaly, second chance algorithm, working-set, thrashing, prepaging, demand paging segmentation, fragmentation(internal and external), swapping, contiguous memory allocation (fixed and variable), First fit, best fit and worst fit algorithms, disk latency, seek time, SEEK, LOOK, C-SEEK, C-LOOK, FCFS, SSTF (shortest seek time first) alogorithms Architecture cache, multiple level cache, SRAM, DRAM, tag, prefetching, cache line Gray code, Hamming distance ALU, pipelining, read-after-write data hazard, pipeline stall, out of order execution, speculative execution, branch prediction Of course the list can keep going... Anyone thinks a list like this is useful? Please add terms that you encounter while solving problems. George