1212448500 =>компютри Monday, June 2nd 2008, 16:15:00 (GMT -7)
If I am right (please correct me if I am wrong!), that would be Tuesday, June 3rd 2008, 1:15:00 (GMT +2) in my time zone or Tuesday, June 3rd 2008, 4:45:00 (GMT +5.5) in India.
I bet we won’t have much sleep that night 
(via GoogleBlog)
Popularity: 1% [?]
Geek. 
As posted in my blog, I believe that if 1212448500 represents the “time in seconds since the Epoch (00:00:00 UTC, January 1, 1970)”, it points to 00:15:00 in UTC+0/GMT+0 which is 1 hour less than your prediction.
How did you calculate that date and time?
Best regards
Thanks Mário.
I used this online converter. Faster than start my webserver and write two lines of code
Isn’t CEST (Center European Summer Time) UTC+2 hours? @see http://www.timeanddate.com/library/abbreviations/timezones/eu/cest.html.
Well, I will be up at midnight and at 1am
Thanks again
It appears I forgot the Summer Time adjustment for Western Europe which started March 30, 2008 and will end October 26, 2008.In that case it means that when I calculated the date/time as 00:15:00 it was in GMT+1.
You are correct CoCaman!
Hi Mário
I guess I will see you tonight at 01:15am CEST (GMT +2)?
Good luck with the puzzle
I’ll be there too =)
done! was quite funny ^^
nice one cr4y
I went to bet at 2:30am. My Eclipse always dies with so many prime numbers… but I keep trying!
How did you solve it? Scripting or programming?
I used a pre generated list of the first 100.000 primes (think this should always be enough) and a little c# programm for summing up. get’s the result in under 40sec =)
Hi everyone!
Congratulations cr4y!
I tried half yesterday half today solving this… I’m not sure if I beat it but I have the feeling I haven’t solved it and I have to “come back in 11 hours to see how” I did.
I spent many hours thinking I had too few PI numbers and that the program was taking too much time to calculate them… Then I thought I had made mistakes with the reasoning behind the sums…
This one is tricky!
hrhr thx =) if you post your sum counts i can validate your result against my code!
Thanks cr4y. Could you send me your counts too?
[11, 83, 183, 631, 1037] to avoid any possible *spoilers* my solution has 7 digits, starts with number 7 and ends in 3.
Do you confirm this?
Hi Mário,
i just finished my solution and tried it out with your values. Now i am happy that i seem to have the same result
.
My solution computes its primes itself and takes about 0,56 sec at all
. Too bad, i could not do it yesterday night 
Greetings
75*4**3 =) Confirm!
Your question: [7, 83, 245, 845]
Your answer: **717*3
Your answer was: Correct =)
Yes cr4y,
5*****3
I’ve just received the confirmation from Google : “Your answer was: Correct” after 11 long hours waiting for the result =)
By the way, I’ll be blogging in the next couple hours with source code included of my solution in C++. It does pre-calculation of PI values in runtime and has solved both problems ([7, 83, 245, 845] and [11, 83, 183, 631, 1037]) in 4 seconds.
Thanks for your input guys!
Maybe I’ll join the fray… first off, on the time thing… launch python and type:
>>> import time
>>> time.ctime(1212448500)
‘Mon Jun 2 16:15:00 2008′
For the prime number solution, I also did it in python… didn’t get started until 8 hours after the fact, but I did it quickly. Output of mine for Mario’s set of [11, 83, 183, 631, 1037] shows:
Generating list of primes…
Done.
Took 3.477478 second(s) to generate.
Computing…
Took 0.312693 second(s) to compute.
Our Prime = [Omitted]
The [7, 83, 245, 845] set took 0.222756 second(s) to compute… python lends itself really nicely to this problem. I was given the set of [3, 37, 397, 589], which took about 0.4s to compute (all on my MacBook Pro).
So yeah, I generate an array of primes and then compute based on that array. Generating the array takes most of the time.
I also confirm what uzi is saying about the time it takes to compute the primes.
$ ./sumprimes
Precalculating…
348513 primes pre-calculated
Took 3.01764 seconds <- the time it took to calculate primes up to the integer 8000000
Result: ******* [ommited]
Took 0.0612206 seconds <- the time it took to calculate after primes were generated
Which algorihm do you use? I get achieve this performance with the {11, 83, 183, 631, 1037} example, using Java on a Core Duo 1,8 Ghz:
Precalculating primes up to 8000000
finished in 0,433937 sec
found 539777 primes
Calculating lists
finished in 0,038925 sec
result: *******
total time: 0,502695 sec
Great results Christopher! When I saw your results now I felt kind of surprised because they are way much faster than mine. Now I noticed a bottleneck in my program was that I was using the bigint library from the beginning (to avoid integer overflows and all). I changed it to 64 bits and it turned out to:
$ ./sumprimes
Precalculating…
348513 primes pre-calculated
Took 0.221172 seconds
Result: 7514623
Took 0.00250829 seconds
I’m using the Sieve of Erastothenes to generate the primes as you check from my blog (solution in C++).
Ooops… I forgot to remove the result from my previous post.
I solved it using just unix tools:
http://www.catonmat.net/blog/solving-google-treasure-hunt-prime-number-problem-four/
Peter
Has anyone heard about the results of the treasure hunt? I guess they have forgotten about it.
Hi there,
Not really. I remember that a few weeks after the puzzles where published, I was googling and checking GoogleBlog but I found nothing about the results of GTH. I guess Google contacted the winners directly but it would be nice if they had at least published the results not to speak about (at least) their official comments on the possible solutions for the 4 puzzles… Of course it shouldn’t be difficult to find thousands of solutions from folks like you and me published all over the Internet, but I would like to see how they would expect the problems to be solved and also to speak about the less obvious solutions that people proposed.
If anyone finds where (if) Google has published the results, please let me know.