Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

The definitive proof that defragmenting an old SSD (wo/TRIM) can work..?

M

Mac32

Suspended
Original poster
Nov 20, 2010
1,263
454
I don't know if this is the right forum or not, but since I did this on a MacBook Pro, I'm posting this here.

I have an MPB 2009 model with Snow Leopard, with a Corsair P128 SSD. This is an early model without Garbage Control and Trim.
These last few months I've noticed that some processes started to hang occationally, for instance watching flash videos on youtube and playing computer games like Half-Life. I noticed the same problem on two different partitions on my P128 drive, both in Windows 7 and OSX 10.6.

I bought iDefrag, and did a full defrag, and lo and behold - the stuttering and hanging processes (video, gaming) is gone!

So, it seems it's just as I've always believed - defragmenting an SSD might actually make sense.
HOWEVER, if you own a newer model with TRIM and GC I would NOT try defragmenting. Also the newest models have an insane IO/per second capability, so these models doesn't compare very well to my P128. For older SSDs however, this seems to work as far as I can see!
 
Last edited:
What do you mean by "It makes sense"? From what you've mentioned... it doesn't make sense...

Could it be, that by defragmentating, it just resets the SSD's controller so it knows where the files are better then before?
 
ayeying said:
Could it be, that by defragmentating, it just resets the SSD's controller so it knows where the files are better then before?
Click to expand...

Nah. What idefrag does, however, is optimize some often used files (called hotfiles). If you got a badly fragmented file system, this might make a small difference.
 
ibd said:
Nah. What idefrag does, however, is optimize some often used files (called hotfiles). If you got a badly fragmented file system, this might make a small difference.
Click to expand...

um theres no such thing as a 'badly defragmented' file system on ssd. fragmentation happens on platter drives where sequential reading/writing happens. thats not applicable to an SSD at all.

in short, don't listen to the OP. all he did was basically add extra wear and tear to his ssd drive by defragmenting for no particular reason and most likely made it worse.

even idefrags OWN BLOG says this:

http://www.coriolis-systems.com/blog/2009/04/real-world-fragmentation.php

"1 For avoidance of doubt, we strongly recommend that you don’t try to defragment your SSD-based volumes. The fragmentation issue on SSDs is internal to their implementation, and defragmenting the filesystem would only make matters worse."
Click to expand...
 
Mac32 said:
I don't know if this is the right forum or not, but since I did this on a MacBook Pro, I'm posting this here.

I have an MPB 2009 model with Snow Leopard, with a Corsair P128 SSD. This is an early model without Garbage Control and Trim.
These last few months I've noticed that some processes started to hang occationally, for instance watching flash videos on youtube and playing computer games like Half-Life. I noticed the same problem on two different partitions on my P128 drive, both in Windows 7 and OSX 10.6.

I bought iDefrag, and did a full defrag, and lo and behold - the stuttering and hanging processes (video, gaming) is gone!

So, it seems it's just as I've always claimed - defragmenting an SSD might actually make sense.
HOWEVER, if you own a newer model with TRIM and GC I would NOT try defragmenting. Also the newest models have an insane IO/per second capability, so these models doesn't compare very well to my P128. For older SSDs however, this seems to work!
Click to expand...


Can we have this **** lying banned?
 
@mac32 let the haters hate. i love defrag as well. i do it daily cause it keeps my computer fast as hell. do you alread do it daily? i am thinking about letting it automaticly run every hour. i heard it extends ssd lifetime as well;)
 
gbit said:
@mac32 let the haters hate. i love defrag as well. i do it daily cause it keeps my computer fast as hell. do you alread do it daily? i am thinking about letting it automaticly run every hour. i heard it extends ssd lifetime as well;)
Click to expand...

You are very funny, aren't you? :rolleyes: Obviously running iDefrag fixed an annoying issue I've had for many weeks. I was just about to buy a new SSD, but now my MBP seems to be running much better.
So, again, iDefrag made a big difference for me, and my old SSD without TRIM and GC.

Btw, one thing that really annoys me are pimple-faced "experts" believing that they are they are always right.
 
gbit said:
here is the problem ;)
Click to expand...

:rolleyes: Obviously this worked very well for me, so it might work for others too, but you know best, right? Btw, have you ever tried defragmenting an older, slower SSD model that's been used for several years?? Didn't think so.... :rolleyes:
 
Mac32 said:
:rolleyes: Obviously this worked very well for me, so it might work for others too, but you know best, right? Btw, have you ever tried defragmenting an older, slower SSD model that's been used for several years?? Didn't think so.... :rolleyes:
Click to expand...

According to my friend, The SSD got even worse after a defrag. It was one of the first SSDs I ever bought, used it for a while, sold it to him. Surprisingly, it still works today because I bought it in 2009 but the speeds are just as fast as a 7200RPM drive today.
 
Defragging a SSD can make it more organized from an OS standpoint. Possibly useful if you ever want to do diagnostics if there is filesystem corruption.

Of course, the defrag operation introduces risk that a bug in the defragmentation you increase the risk of corruption.

Now, at a lower than OS level, the drive most likely has increased its fragmentation. With wear leveling, it scatters the filesystem blocks around the drive.
 
Mac32 said:
If anyone should be banned it's you. I'm just reporting my findings.
Click to expand...

Your "findings" are useless, ill-advised and not even recommended by the people who made the program you're talking about. Just stop.
 
I know this seems to have gotten some people hot on here (I'm not taking any sides, either), but this quote is accurate:

um theres no such thing as a 'badly defragmented' file system on ssd. fragmentation happens on platter drives where sequential reading/writing happens. thats not applicable to an SSD at all.
Click to expand...

Not only that, but the way defragmentation works is by moving files from one location on a drive platter to another, then rewriting them sequentially so they're lined up in contiguous tracks on the platter.

SSDs know nothing about platters, BUT, it might be possible that defragmentation is placing them in contiguous blocks of SSD memory, thus making the process of looking them up faster. For example, suppose a data file was written to a fragmented SSD, and lets say, for example, that to do so it had to write it into 500 different blocks of the SSDs memory with each being of different sizes. After defragmentation, I would think that the defrag software would make this more or less a single look up - i.e. start reading at point X in the SSDs memory and sequentially read Y blocks. I would think that **might** happen, but it's only a guess (and it should be taken as ONLY a guess). I find this theory **sort of** acceptable, but how long does it really take an SSD controller to look up and access a sequential block of memory as opposed to, say 500 different chunks. I wouldn't really think the performance change would be that big, if even noticeable - unless the SSD controller itself was abysmally slow.

What I would be more concerned about doing a defrag on an SSD is the fact that the memory cells have limited write counts, and with defragmentation being extremely write intensive, unless the defragmentation yielded literally night and day results, it may be doing little more than shortening the life of a device.

With your SSD being kind of old, perhaps some older, marginal cells were mapped out as bad and remapped into decent cells during a defrag. I would think this might imply your SSD is on the way out.

Just my opinions fella's - not the word of God! :cool:
 
Last edited:
ZVH what you say makes sense to me. When I bought my P128 it had listed two different speeds, one for continuous read/write, and one for random read/write. Doesn't defragmentation come into this picture?

Again, as I've already stated that this should never be done on an SSD with TRIM (or GC), and this means only the oldest SSD models that's 3 years or more..

As far as speeds: the reading speed of my P128 is almost as good as new, but because it lacks TRIM and GC the writing speeds have slowed down quite a bit. But it's easily still better than a regular hard drive... (if you don't know this simple fact about SSDs, you shouldn't pass yourself off as an expert IMO.)
 
Last edited:
If ever there was a moving target on the face of the Earth, SSDs are it!

You should take my comments as educated guesses and nothing more. Why? Drive developers don't always, how should I say it, present facts as they really are. Take S.M.A.R.T. technology. Many of them fail to fully comply with S.M.A.R.T. standards but they advertise them as fully compliant. Another stunt they used was to take EIDE drives and stick a SATA interface adaptor on the drive's controller card and then claim it's a "real SATA drive." It isn't, it's an adapter in front of the drive.

True SATA requires different specs and higher speeds. That means higher speed chips, which cost more. Trying to hide data transfer speeds behind buffer to system data rates seem to be a common trick even though the actual transfer speed on a drive remain unchanged.

I would NOT be surprised if your SSD is yet another kludge where the designers, forced by cost from their superiors who have an unending desire to line their pockets with other peoples money, put together a system using essentially an IDE or EIDE controller which can't properly handle the data rates an SSD is truly capable of ... it would be done with cheap, lower cost (and possibly) much slower chips. Such systems don't need high performance chips, but back then it would still look lightening quick with respect to an actual hard drive. Enough of me being on a soapbox! :eek:

Like I said at the start of this post, SSDs are a moving target. The technology is changing rapidly and dynamically. I cannot comment on your specific (or for that matter) any SSD drive for this reason. However, what you might want to monitor are actual changes in the size of available drive space over time. SSDs, as I understand it, don't do what USB flash drives do, which is, or was, to mark the drive "read only" once enough pseudo sectors have had their write limitations exceeded. I think SSDs attempt a reallocation to a decent sector and then mark the semi-functional sector as "bad." Frequent or constant defragmentation would simply be a way of killing more and more pseudo sectors over time because of the number or write cycles needed to defragment a drive. It's sort of like the way a true hard drive uses to be able to remap bad, slow, or unresponsive sectors by zeroing the drive. If this is the case, then it's kind of a crap shoot: What do you want ... drive space or performance?

Like I said before, just my opinions fella's - not the word of God!
 
Thanks for some interesting and thoughtful comments! I guess the moral of the story is: You might try defragmenting an older SSD without TRIM at your own risk, but you should never attempt it with a newer model that has TRIM (and GC).
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back