And even if the files on your chip were fragmented it wouldn't hurt one bit. It was about 20 years ago that defragmented files slowed down a computer, but with modern processor speeds it stopped being an issue 10 years ago. So movies will record onto even the slowest chips without a problem. Since your camera is so old and slow you probably don't do any movie recording so you are not in a very good position to be advising people about how slow speed cards can't record movies.
?!
No. Although some of the information here is right, a lot of it is wrong.
Fragmentation absolutely effects modern hard-drive based computers, and it can seriously effect performance. Why you would think that stopped being an issue 10 years ago is beyond me. In fact, the increase in processor speed makes it more of an issue today than ever because the faster the processor is the sooner it runs out of data to process and the sooner it will be waiting for the hard drive.
But that's a digression because it doesn't apply to cameras.
Cameras use flash memory, and for the most part, solid state memory does not slow down with fragmentation, at least not to the extent that hard drives do.
Hard drives have physical heads that must be moved around. It takes about 8ms to move a hard drive head from one place to another, and if the head must be moved about 100 times to read a single file because of bad fragmentation, then that adds .8 seconds to the time it takes to access that file.
Flash memory, on the other hand, has no moving parts, and in general has seek time. The seek time from flash results because entire blocks must be charged for reading/erased for writing at a time. Thus, on a memory card, you would ideally like for the card to be completely blank so you can write sequentially to it. Otherwise you may have to do "read/update/write" cycles to the memory which are slow.
The bad news is that formatting a flash card doesn't completely return the card to this "erased" state, and you cannot de-fragment a flash card the way you can a hard drive.
Why? Because of wear-leveling. Internally, the flash card assigns blocks based on which blocks are currently available and have been written to the fewest number of times. Formatting a card will allow you to write complete blocks at the logical level, but at the internal level those blocks will still be scattered all over the place.
There is some improvement to be had by formatting, because in general the card will be doing the fewest number of "read-update-write" cycles necessary to write data, and no more, but you will never get back to out-of-the-box performance.
SSD hard drives suffer from this same problem. Look up the "trim" command for SSDs.
The good news is that the issue is almost completely moot for cameras. The added overhead is relatively minimal anyway.
As a side note, I don't trust any of the existing CF benchmarks because of this.
The correct way to benchmark a CF would be to intentionally fill the card with fragmented files many times to "scramble" the internal map, THEN format the card and THEN test it. Otherwise you can end up comparing a virgin card to a wear-leveled card which is a dramatically unfair comparison, as erasing a flash block takes MANY times longer than writing to an empty one.
That's why I take the 533x PhotoFast superior result when compared to the 600x Transcend with a grain of salt (on the eos 7d.)
