I don't think Google will make that public; they are likely to use a much more complex version of Porter stemmer. They might also be using lemmatizers in addition or as alternative to stemming!
Google is even parsing parts of the English web, so they don't have to rely on 1980's technology like Porter Stemmer. Sure that they have something better than that.
a) Porter is really 1960's (or maybe 1970's) technology ... I can give refs if you want
b) I think Google in most languages uses little or no stemming - stemming is a recall enhancement technique, and google is a precision orientated engine. They may use morphologically based query EXPANSION if they find very few (or zero) results. In any case they always are running at least two engines for each language and change them (on both the index and query sides) very frequently - so there isn't really an answer to your question.