Is it true that writing to a disk drive having lots of empty space is faster
than writing to one having less space? Suppose I have 1 GB of data to write and
I have felt it is faster to write this data on to a disk drive having a space
of 10 GB than on a drive having space of 5 GB? But I’m an electronics guy.
Theoretically, the given amount of data should give the same amount of time to
be written irrespective of the remaining space on the drive. Of course, given
the empty space on the disks are greater than the size of the files being
written to it.
In this excerpt from
Answercast #18, I look at the time it takes to copy files onto a disk with
lots of space compared to copying the same files to a more full disk.
Slower on a full drive?
This is actually a really interesting question. I’ve never thought about this way. The surprising answer is maybe. It could be faster on an emptier drive.
So the concept here is that you have a drive. It has say, 20 GB of free space and you write 5 GB of data to it and it takes “X” amount of time. You go to a different drive, repeat exactly the same kind of a thing except that the different drive has only 10 GB of space available. The 5 GB copy is just exactly the same copy you did before.
It feels like it takes longer.
It’s about fragmentation
This isn’t really about the amount of free space on the drive. What it kind-of boils down to is (of all things) fragmentation.
A drive that is full, or close to being full, is more difficult for the defragging tool to defrag. In fact, sometimes, defragmenting of many files simply won’t happen if there isn’t really enough room.
What happens then is on a drive that is more full than another, files tend to be more fragmented. Conversely, the free space into which you’ll be writing your new file is also more fragmented.
What that means is that the drive head, the physical head that runs around the hard disk platter as you write to your drive, is doing a lot more movement as it tries to fill up all these disconnected holes with the five gigs you’re putting in.
If, in both of these cases regardless of the amount of free space, your drive is thoroughly defragmented (so that all of the free space has been collected up and lives together, basically as one big empty block of free space) then I would expect the copy to take roughly the same amount of time: regardless of how much free space there is.
More fragmentation in a full drive
I think that there is a coincidental correlation between the amount of free space and the amount of fragmentation that is probably apparent on the drive.
And as the science guys like to tell us, “Correlation is not causation.”
It’s not necessarily the free space that means you are being slower; it’s that underlying defragmented nature of the hard drive that could be slowing you down.
So what I encourage you to do is repeat your test; but this time, defrag the drives first to make sure that they’re as defragmented as possible. I’m guessing you’re going to find that the copy operation will be pretty consistently the same amount of time, regardless of how much free space is on the drive.
Next from Answercast 18 – Why are my two anti-virus tools giving different results?
8 comments on “Is it faster to write to a drive that has lots of space than to one with less space?”
Wouldn’t it also depend on what part of the drive the data was being written? The outer edge of the drive should be faster than the inner. But doesn’t data get written to the inner part of the drive first and expand outward?
Hi Leo, I seriously did not think this could be the answer when i faced this issue. I did not even think on these lines. Nice answer. I will verify it for you by defragging the drives and copying stuff again.:) And one other question, A drive should not be fragmented, Why does not the copying process copy the data in such a way that fragmentation does not occur? I mean why don’t they copy the data serially without hatching “disconnected holes”? Does it take more time doing it that perfectly?
When the drive is new or defragmented, the computer copies the new files in sequence to the hard drive, but after a while when you start deleting files which happen to be in the middle of all of your written data, it writes the data into the new free space created by the deletion. As it doesn’t always fit perfectly, the data is then written to different free areas on the drive. That’s what fragmentation is, in a nutshell.
The inner part of the drive is faster. That’s why the data is written there first.
I would like to know the comparison results of writing to disks before and after defraging. I have a lot of free space on my disks or I’d do it meself.
If there is enough space to copy the file as one contiguous file, then outer edge -or- in the middle, doesn’t reall make much of a difference. Yes it will take a fraction longer to get to the middle, but once it is there – the head wouldn’t move any more or less. If the head had to move all over the disc, then it would make more of a difference. For instance, you carry a box of books to the shelf right next to the door, set down the box and fill the shelf. That takes a certain number of minutes. Instead you go two shelves over, set down the box and fill the shelf – filling the shelf took the same amount of time, the extra steps to that shelf took more time, but only once. If you have a fragged shelf system, you set down the box, and have to walk back and forth to put books in different shelves – that is where it takes a lot longer, because of all the movement you have to do in between moving books from the box to the shelf.
if u want that there must b same time in both case them defragment ur hard drive time to time ….
Back in the days of the XT Turbo and even the x286, I used to religiously defrag my hard drive, based upon the still prevalent perceptions of speed. In all of that time, I never saw any difference in speed nor performance. By the time the Pentium came out, I’d discontinued the practice. Even recently, I’ve read articles in various PC magazines about tests done on fragmented drives vs defragged drives, and any variation in performance has been negligible; on the order of a couple seconds out of 10 minutes. Yet, the myth persists that Defrag is a MUST to be periodically performed. Seems almost like praying to some ancient god because that’s what our long forgotten ancestors once did. In other words; superstition.