Super GM's have obviously memorized many variations; however, it is humanely impossible to memorize all the variations of any one opening. If someone is an expert in a certain opening, it means he knows all the themes and motifs for both sides: for example, in the Sicilian, it is a well-known manuever to play a6-b5 for queenside counterplay, or in the Caro-Kann, to play a5-a4-a3 to start an offense on White's king. So Giri knows all the tweaks of the Gruenfeld, and how to accomplish those goals. Hope this helps.
Memory of a SGM

I doubt Giri knows as much about the Gruenfeld as either Anand or Topalov
who are on another level. Anand's mistake was just that - an anomaly.
Although, the greats play the same openings as amateurs the meaning/understanding behind
the moves is often much different.

Super-GMs certainly memorize lots of variations, but I guess they also forget and have to reharse what they have learned from time to time.
GM Nakamura mentioned in an article that he learnt like 1000 moves of theory before every game, and Bareev in the book 'From London to Elista' explains how Kramnik used to rehearse all the lines his team has been working on before a WC game.

Unfortunately, at every level the person's analyses exceed his/her memory capabilities. One can memorize three screens of lines and has ten of them; other can keep in mind 30 screens, but has over 100 of them in the database.

Giri cannot be called a supergrandmaster. I think SGM is considered something like 2750+. Giri is not even above 2700.
Anand had only mixed the move order. He is probably the best at openings in the world at the moment, together with Topalov. They do have amazing memory, but that's also partly something they have developed by working a lot.

When I first saw the title of this thread, I actually thought it was a memorial thread for some SGM who had passed away.
Expected

Giri cannot be called a supergrandmaster. I think SGM is considered something like 2750+. Giri is not even above 2700.
2600+ is a SGM

That's why 2750 is an appropriate level. The super cream of the crop, the people who will lead the big tournaments - they are Super Grand Masters, not the group of people below them.
LOL, 2600+ is a Super Grand-Master? Even in the professional circuit, at the top-level the people rated 2600 are literally unknown in the world - ranked about 200th!
I think the ranking criteria for Super Grand Master should be by World Ranking rather than by ELO, as there should pretty much be the same number of SGMs in the world now as there were, for instance, in Fischer's time. I would put the barrier somewhere ranging from Top 7 to Top 10 (currently around 2750 in ELO).

2600 is actually as normal a GM can get. There's nothing super about it. Only discussion I'd make is if 2700+ could be called super gm, which I stand 2750 should be the limit.
And then comes the "elite" level, Anand, Kramnik, Carlsen, Topalov, Aronian.

Maybe when he was working with Carlsen, but at the moment, I seriously doubt if Kasparov would measure up to Anand's preparation. This needs full time dedication, which is clearly lacking for Kasparov since his retirement.
Giri may become a SGM in time, but I agree with Philidor here...
2600 is a standard GM as far as I'm concerned. I mean, aren't a great % of GMs 2600?
As far as SGM goes... I could live with the 2700+ argument, but I also think it should be 2750+
And to take it one step further: I think FIDE should formally recognize SGM as an obtainable title.
It's pretty unfortunate, though true, that you can call someone a "standard GM", because that kind of destroys the prestige. Wasn't the GM title recognized at one point as a title given to someone whos playing strength is enough to have real chances at becoming world champion? Now that's no longer the case.

Giri may become a SGM in time, but I agree with Philidor here...
2600 is a standard GM as far as I'm concerned. I mean, aren't a great % of GMs 2600?
As far as SGM goes... I could live with the 2700+ argument, but I also think it should be 2750+
And to take it one step further: I think FIDE should formally recognize SGM as an obtainable title.
It's pretty unfortunate, though true, that you can call someone a "standard GM", because that kind of destroys the prestige. Wasn't the GM title recognized at one point as a title given to someone whos playing strength is enough to have real chances at becoming world champion? Now that's no longer the case.
One of the original GMs was Frank Marshall and he had no chance to be WC.
When they say "He is an expert of ........ defence." Like for example SGM Arish Giri is an expert of Grunfeld. Does it mean he remembers ALL the ideas in the opening that has been played from the database? Is their memory so good? Actually in the Anand-Topalov WCH in game 1 where Anand lost it was because he forgot how to respond to one idea in Grunfeld where Anand had Black. But then again Anand is not an expert of a certain defence I guess... it`s like he knows a lot but does not stick to just one or something?