Does READ speed deteriorate when SSD gets full?

ondra30

Honorable
Sep 2, 2013
6
0
10,510
Hi,
I already know that when SSD gets full, it "slows down" since there's not enough space to correctly perform GC, but does this "slow down" affects read or write speed?
My 238GiB SSD is primarily used as read-only drive. It's filled with 237GiB of data and won't be written to in the next 3 months. HD Tune showed average read speed 460MB/s for 1MB block size, which is still OK, but it seems to me that 4K read/access times got a bit worse, now it's 75MB/s and 0.138ms access
 

MrTomnus

Honorable
Sep 2, 2013
13
0
10,520
The amount of data stored on the drive shouldn't affect performance. The principal of an SSD is that there are no moving parts so all the data on the drive can be accessed in the same time, regardless of its location on the drive, unlike an HDD. This would mean read speed isn't affected. By the same principle I would assume writing to the drive would't be slowed due to a large volume of data, but I'm not a huge expert on them.

One thing I do know is that an SSD itself deteriorates with more reads. Your drive is quite full so I would assume you'd clean it around regularly to keep your file system tidy, which could be the cause of the slow in times. But still, 0.138ms is 10,000ths of a second, which is still incredibly quick access in my book, and won't be noticeable.
 

ondra30

Honorable
Sep 2, 2013
6
0
10,510


I already tried removing a lot of things, but this is what I ended up with. I should have probably said from the start, this SSD is not my system drive, there's only user-generated content.
However I have to disagree, SSD does not deteriorate with more reads, but with more writes. Read endurance for SSD is unlimited, write endurance is limited to about 10,000 rewrites for SLC, 3,000 for MLC and 1,000 for TLC
 

MrTomnus

Honorable
Sep 2, 2013
13
0
10,520
Haha, you seem to know more than me. :) But it appears to me that you've answered your own question. If the read endurance for an SSD is unlimited, then I would have thought there shouldn't be any reason why the speed at which the data is copied would be the same. Again I'm not an expert, but if that's not true then it would depend on if the data takes one or two lane traffic, so to speak. If written data takes a different path to data that is read from the drive, then the read speed shouldn't decrease. But if the written data moves with data to be read, then the path would be "worn down" from all the writes, making it harder to read, no?

Think I'm getting a little out of my depth :D