Would defragging an SSD make a difference?

[hat]

As we all (or at least most of us) know, defragmenting came about because stuff tends to get scattered all over the place on a hard drive; for example, if you were loading a Crysis map and the files for that map were scattered all over the drive, it would take much longer to load than if it were all in one place because the drive would have to look for the files, which greatly slowed things down with mechanical hard drives because the read/write heads only move so fast. However, a SSD hard drive is basically a chunk of flash memory with no moving parts, so it isn't hindered by the things that crippled the speed of mechanical hard drives... so I was wondering if defragmenting a SSD hard drive would make a difference since it doesn't have any moving parts??

 

[oily_17]

Defraging SSD's is bad for them..you dont need to defrag it.

 

[human_error]

Never defragment an SSD. They way they work is by splitting data up all over the place, so when it is needed they can get as many bits of data from as many memory chips as possible - this makes the operation more parallel which is where ssds get tehir speed from. Defragmenting won't actually defragment the data on an SSD anyway, it will just move it around which takes read/write cycles out of the ssd unnecissarily shortening the SSDs life whilst gaining nothing.

Also the computer doesn't know where the data really is on an SSD anyway - the SSD's on-board controller is the only thing that does, and it tries to split the data up to get fast, parallel operations when retreiving the data back.

 

[hat]

That went right on my train of thought.

 

[RejZoR]

Actually SSD's get fragmented just like traditional HDD's and they start to suck in performance just as well. So the myth about SSD's not losing performance when fragmentation increases just aren't true.

 

[human_error]

Actually SSD's get fragmented just like traditional HDD's and they start to suck in performance just as well. So the myth about SSD's not losing performance when fragmentation increases just aren't true.

Source on this info please - this seems entirely wrong from all the articles about SSDs that i've read (especially since there is no time cost for getting data from different parts of an SSD, where hard disks lose time getting data which is why fragmentation on HDDs is bad but is good on SSDs).

 

[oily_17]

Actually SSD's get fragmented just like traditional HDD's and they start to suck in performance just as well. So the myth about SSD's not losing performance when fragmentation increases just aren't true.

Well this is a quote from OCZ forum on their SSD's

Drive Defragmentation:
DO NOT DEFRAGMENT your SSD, it is not necessary. Wear leveling technology purposely "fragments" the data to ensure the consistent life of the drive.
VISTA has an Auto defrag feature that is enabled by default (defragging is on a schedule) this should be disabled after your drive is installed.

And a quote from another source on SSD's, which can explain SSD performance loss, which is actually due to defraging.

The SSD is silently moving data to other flash cells without Windows or anything else ever knowing. Now here comes the trick: it has to remember where all these 'mappings' correspond to. For every I/O this 'list' needs to be referenced, to see where the data really is being stored. But with defragmenting this list can grow enormously, and this starts to hamper performance.

 

[shevanel]

informative thread. thanks for the info

 

[Stearic]

A conventional defrag will not help the random read and sequential read/write performance of an SSD much. However, free space fragmentation can reduce the random write performance of SSDs* So free space consolidation can help to minimize the chance of random writes at both the logical disk level as well as the SSD's physical cell/block level.

It's important to remember that the 'internal fragmentation' i.e. wear levelling of an SSD is a function of the drive's firmware and has nothing to do with the filesystem fragmentation which occurs at the logical disk level. The SSD's hardware is abstracted from the filesystem i.e. the file system does not know whether the drive is a SSD or a HDD**.


*Which is their weak spot anyway because of the mismatch between the logical cluster size and 'physical block' size (4 kB vs 512kB) requiring frequent erase-rewrites for entire physical blocks.

**Maybe different for Win 7 since it is said to have the ability to detect/recognize SSDs and disable defrag accordingly.

 

[ERazer]

Actually SSD's get fragmented just like traditional HDD's and they start to suck in performance just as well. So the myth about SSD's not losing performance when fragmentation increases just aren't true.

what the heck r u talking about? do u even own SSD? wers ur source of info?

 

[oily_17]

A conventional defrag will not help the random read and sequential read/write performance of an SSD much. However, free space fragmentation can reduce the random write performance of SSDs* So free space consolidation can help to minimize the chance of random writes at both the logical disk level as well as the SSD's physical cell/block level.

It's important to remember that the 'internal fragmentation' i.e. wear levelling of an SSD is a function of the drive's firmware and has nothing to do with the filesystem fragmentation which occurs at the logical disk level. The SSD's hardware is abstracted from the filesystem i.e. the file system does not know whether the drive is a SSD or a HDD**.


*Which is their weak spot anyway because of the mismatch between the logical cluster size and 'physical block' size (4 kB vs 512kB) requiring frequent erase-rewrites for entire physical blocks.

**Maybe different for Win 7 since it is said to have the ability to detect/recognize SSDs and disable defrag accordingly.

Also, IIRC, with Win 7 using the TRIM function, free-space reclamation is done automatically when you delete a file.

 

[RejZoR]

What the hell i'm talking? Then why the hell read/write performance drops over time when files get fragmented (by fragmented i mean fragmented files and fragmented free space with non cleared cells)? It's not something i made up, it's something i've read in many SSD reviews...

 

[TheMailMan78]

Defraging SSD's is bad for them..you dont need to defrag it.

I did not know this. Thanks for the enlightenment.

 

[Cold Storm]

Man, I gotta say, I love my new ssd that I got from Bogmail (Intel X25-M "80")... Best part.. came in installation instructions.. You know the other best part... If you read it... Page by page... It states the following at the end.


Post Installation Considerations


SSD's do not need to be defragmented like Hdds. Turn off any disc defragmenter to prevent unnecessary wear to the SSD from Microsoft* Windows Vista*, use the following instructions to turn off the disk defragmenter:
1. From the start Menu, clock on All Program > Accessories > System Tools > Disck defragmenter

2. Uncheck the "Run on a schedule" option and click OK


For more information about Intel SSDs including tool downloads and drive optimization suggestions, go to http://www.intel.com/support or to http://www.intel.com/go/ssd



Sorry for the first part.. I just had a little hair up my butt when I started this post..


Now, from what I have heard via SSDs is that it's not good to defrag at all.. I personally never read a review on the issue, and from the people I know who write them... They say they heard the same.. Not a good thing to do so..

 

[erocker]

What the hell i'm talking? Then why the hell read/write performance drops over time when files get fragmented (by fragmented i mean fragmented files and fragmented free space with non cleared cells)? It's not something i made up, it's something i've read in many SSD reviews...

You still should never defrag an SSD. Personally I've never seen a SSD review that suggests it.

 

[MN12BIRD]

He's right in a way. SSD's should write info spread across different cells from one write to the next to allow the cells to last longer. The problem with SSD's is the way they have to write blocks. They can't write a bit here or there. If a small update is made to a small file it has to copy an entire sector to another area on the drive, delete a block and rewrite an entire block, delete the copy. So if the block size is 64K then everytime a file is changed only by 1K it can read 1K but has to rewrite 64K. The data in a way does get fragmented and the performance of SSD's does in fact get lower with time. But a traditional defrag is not the answer. You need the TRIM command with an OS and firmware on the SSD that supports it. This restores the speed of an SSD or at least prevents it from slowing down.

I do not own an SSD, nor have I ever used one. So this is only my basic understanding of them.

For more info look up the "TRIM command" and make sure your SSD and firmware updates support it.

 

[Cold Storm]

Also to add to this debate of things...Just installed win 7 on the ssd.. Intel X25-M... I had no problem what so ever installing it, and win 7 automatically turned off all defraging agents on ALL drives.. Where as when it was installed via my WD Black 500gb, t was automatically defraging on their generic time.. Wed 1am.. on both the main and storage drive.

 

[RejZoR]

There is no need to own an SSD to know NAND flash technology. We all have USB drives and memory cards that work on the exact same tech. Just with smaller capacities and lower speeds.
But it acts just the same.

 

[human_error]

There is no need to own an SSD to know NAND flash technology. We all have USB drives and memory cards that work on the exact same tech. Just with smaller capacities and lower speeds.
But it acts just the same.

You're confusing file fragmentation (which is fine on an SSD) with the loss in performance due to old data not being deleted in cells resulting in slower writes as the write cycle has to deal with the old data first (which is not called fragmentation).

To get the old data deleted from the cells you don't defragment (which is to do with file fragmentation increasing the amount of time it takes to retreive a file's data) but you do use a setup which utilises a TRIM style setup where the old data is overwritten with blank space in the memory cells, removing the loss in write performance (and helps reduce the small loss in read performance as well). That's where you get the loss in performance in an SSD (if TRIM is not used), not due to fragmentation.

 

[RejZoR]

Then i made all up and the tests ive seen were all fake. Yeah, that was it... Unless they were automatically trimming empty cells...

 

[Static~Charge]

What the hell i'm talking? Then why the hell read/write performance drops over time when files get fragmented (by fragmented i mean fragmented files and fragmented free space with non cleared cells)? It's not something i made up, it's something i've read in many SSD reviews...

Then post some links so we can see it for ourselves instead of saying "just take my word for it".

 

[btarunr]

Are sequential and random read speeds on SSDs exactly the same? If not, defragging will make a difference.

 

[DrunkenMafia]

This is a simple case of confusion between trimming the drive and defragging it.

Conventional defragging is useless on SSD's.

Trim is helpful and will keep the performance of the drive tip top..

 

[xfire]

Defragging involves a lot of read/write operations. It significantly reduces the life of the SSD.

 

[douglatins]

I am absolutely sure everyone already responded to that, but i though it was common sense by now