Bitmap Sheets - tiled or single row - which is better?

Question

For alot of my animated sprites, I wrote a little program to read in all the bitmap data, crop the images and output a single sprite sheet of all the bitmaps, plus the "array" data needed to find and copy out the specified frame.
This has proven to save both memory and preformance when dealing with large numbers of frames
However, when I look at several example of how other people do sprites sheets, their final sheet is always tiled
So if I have an animated sequence of 60 frames, (assuming that each source bitmap is, 50x100 and there is no cropping to be done) I would end up with something like this:
Bitmap: Size = 3000x100;
TopLeftCornerArray = [[0,0],[50,0],[100,0]...[2950,100]];
Is there any difference between doing that, instead of creating a sheet bitmap, that was say 10 frames per row over 6 rows giving me a bitmap that was 500,600?
I mean, both bitmaps are made up of 300,000 pixels regardless...
I was just wondering if there was a good reason for not putting them all on a single line.

Answer 1

I am "gridding" it, for the most part, as 50% of all my images are blockish anyway, so there is no real "cramming" needed.
The rest are characters, and unfortunately, due to the way they were rendered and created by the artists, (which was originally build for director) and they had wonky registration points, writing my own allowed me to programmatically calucate where those registration points would have been in director and output that data as part of my frame arrays (which I call "center point offset") while still cropping the images to their smallest size.
I am not "cramming" it like in the texture packer is doing, but given the irregular shapes of most of the character frames anyway, there would be some sort of wasted alpha anyway... so while it's true, using a program like that (and I am actually looking into it), would probably result is slightly smaller images in the end, being able to do the programmatic custom stuff outwieghed the advantage of saving a few hundered pixels per sheet.
But just to make things easier to code, I did it in a long single row instead of tiling.
For example, one of my sheets consists of a "spinning coin". every frame is always 56 pixels in height. only the width of the frame changes based on the width of the coin itself. the vast majority of my sources do this. so with 30 frames, so when I built the sheet, I crop the images and I end up with a final bitmap about 1100 pixels in width, and 56 in hieght, and it's packed just as good as any other prog.
From a programmatic point of view, there is no difference between a single or or multiple rows, but given how memory problems for ios are a nightmare, and performance is a pain, I wasn't sure if there was some "unknown and undocumented reason" why tiled vs single row would be better.