Google made a difference when produced search results in less than second. While that required a lot of parallel processing, the results compensated for the trouble.
Some of the resource-intensive queries are those that include advanced operators (like site:, inurl:). Other factors are: personalized search, OneBox results or strange queries that require access to the supplemental index.
In 1999, the average search took approximately 3 seconds. Now most of the searches take less than 0.4 seconds. But I could find a simple query that took 4.22 seconds, admittedly for the second page of results. I wonder what was the problem (it was something temporary, if I repeat the search - it takes 0.28 seconds). Can you find other resource-intensive Google queries?
the reason the search is fast the second time around could be cache?
ReplyDeleteIs this a joke? If you can't reproduce it, it's not a "resource-intensive" query. It was simply caused by a bigger CPU load at that time.
ReplyDeleteYou're probably right, but it's so rare that I thought it deserve to be mentioned.
ReplyDeletethis query seems to take less than a second (0.32 sec):
ReplyDeleteamerica -bush -early -library -bank -roadside -bbc -summit -toshiba -virtual -epson -oxfam -biography -economia -danmark -historic -television -geological -volksvagen -shape -recycles -teach -scouts -democracy -planned -public
i guess such a query must take some CPU in order to be correct.. but it's nothing compared to what the google cluster can handle.
Hey peeps,
ReplyDeleteJust managed to rank up a massive 6.58 on the Goog'o'meter.
Highscore board anyway?
You probly wont get the same results cause the keyword has been cached on their own machines. But anywhoo,
http://img456.imageshack.us/my.php?image=658ag5.png
This must have been resource intensive on their spell checker.
ReplyDeletehttp://www.google.com/search?q=you%27re+either+with+us+or+against+us&hl=en&sourceid=gd&rls=GGLD,GGLD:2006-20,GGLD:en
Working in search engine, I can assure you that after a query is made once, it will stay in the cache for some time, and thus react much faster the next time.
ReplyDelete