Will you upgrade? That is the question...

[crawfish]

We're talking about file systems, so to get this back on track in terms of things we can actually observe, WTH is a "data tree"? Do you mean "directory"? Directories that "go six or seven deep" are not even moderately unusual, nor is "massive amounts of data storage", whatever that means, which probably isn't much considering the small capacities of the SSDs you are likely using.

You have to understand how the underlying filesystem works. It's built upon a computer science concept known as b-trees and linked lists. What can happen, as files are added and deleted to directory entries, is that they get "holes" in them where deleted entries remain and new entries are added onto the end. After a lot of time passes, this makes directory entries quite large and reduce directory scanning performance (because it has to skip over all these old entries). It's true that an SSD will make this faster in many cases as well, but an optimized DE will always be faster than a non-optimized one, regardless of the storage media.

Still, my point is that defragmentation with a good defragmenter can achieve some benefits.. I don't consider them worth the loss in write cycles caused by defragging an SSD, but still they are there... in some circumstances. I don't advocate defragging an SSD, but at the same time it's not as ridiculous as you like to make out.

I understand how file systems work, and as a software developer for over 20 years, I've spent a lot of my time choosing data structures such as lists, trees, graphs, etc from libraries when possible and implementing them from scratch when necessary. As I replied to Poppa Bear's enthusiastic adoption of your terminology, I wanted "to get this back on track in terms of things we can actually observe". We can't really observe the "data trees" and other OS data structures (not that Poppa Bear understood that's what you were referring to - after all, he talked about "massive amounts of data storage, with data trees that go six or seven deep", and I'm sure he meant files and directories), and everything you've talked about is conjecture based on some knowledge of how the OS works. That knowledge can be the basis for forming hypotheses and designing experiments, but it cannot be used to draw conclusions, which is what Poppa Bear did. For all we know, the optimization of OS data structures you mentioned and I acknowledged as "interesting" but probably hard to measure amount to microseconds in operations that take milliseconds or longer to complete; IOW, they could very well be negligible. I think this is very likely given the relative speed of SSDs, the ridiculous speed of modern CPUs with gobs of memory at their disposal, the relatively small amounts of data needed to keep track of files and directories as compared to the amount of data they contain and what is done with it, and the scarcity of legitimate people recommending defragmentation of SSDs.

 

[poppa bear]

I understand how file systems work ... and the scarcity of legitimate people recommending defragmentation of SSDs.

Wow! I'm amazed! You're not just a crustacean, but can actually venture out of your shell and put together a few logical comments instead of just taking cheap shots!

It seems you've become a legend in your own mind, which has clouded your ability to see things clearly. With all your vast experience, and a man who demands facts, what actual hands on tests have you done to check the difference, and where are the performance graphs to verify your results ... Hmmmm? That should give you something to occupy your time for the next few days.

No doubt from my experience, defragging SSDs does not harm, and does improve performance. End of my story. Believe what you want.

Oh, and BTW, as any aeronautical engineer will tell you, polished car and aircraft surfaces reduce the drag coefficient due to air friction, so cars actually do go slightly faster when cleaned. :thumb:

He who came to mock, became humble, and stayed to learn!

Old Poppa Bear adage! LOL!

 

[crawfish]

I understand how file systems work ... and the scarcity of legitimate people recommending defragmentation of SSDs.

Wow! I'm amazed! You're not just a crustacean, but can actually venture out of your shell and put together a few logical comments instead of just taking cheap shots!

It seems you've become a legend in your own mind, which has clouded your ability to see things clearly. With all your vast experience, and a man who demands facts, what actual hands on tests have you done to check the difference, and where are the performance graphs to verify your results ... Hmmmm?

No doubt in my mind defragging SSDs does not harm, and does improve performance. End of my story. Believe what you want.

You familiar with the saying, "Extraordinary claims require extraordinary evidence?" Here's a free clue: You're the one with the extraordinary claim, and the dancing elephant is not extraordinary evidence.

P.S. Despite what my friend says about washing cars, I've never tried to test whether my car drives faster after I wash it. I've addressed several of your misconceptions and odd statements and explained why I don't believe you over several messages, which you've either clipped into nothing, ignored, or dismissed as simple mockery. You replied to my car story with the self-contradictory statement, "Your cynicism, mockery and ridicule don't warrant a reply. ", and unsurprisingly, you failed to see that it anticipated your latest reply.

 

[poppa bear]

You familiar with the saying, "Extraordinary claims require extraordinary evidence?"

You don't get it do you! I'm not interested in trying to "prove" anything to anybody, and especially not to a self-proclaimed expert and cynic like you. I'm simply stating my experience and my interpretation of the facts. Up until your last post your only reply was to mock and ridicule.

Sounds like you're the one trying to prove something. And as I said before, since you demand facts, produce the performance graphs of heavily loaded OSs on SSDs before and after defrag. Put your money where your mouth is.

If an SSD HD controller is randomly accessing data from various parts of the HD in parallel, it then has to put them in logical sequential order to read the data as a binary code. If the bytes of data are already in a logical sequential order it eliminates the step needed to put them in order. It can just read it straight off without organizing it. And hence is faster, even if only margianlly so. At least that's what I was taught when doing basic programming while studying electronic engineering. And at the hardware level the reading is still controlled by a series of electronic flip-flop switches in the logic chips.

Re extraordinary claims, a M$ techo told me it was impossible to successfully load Win98 on a hard drive, AFTER loading XP, so they both work. Well, guess what? I did it. And M$ didn't want to know how! So much for certainty in the world of PCs. Besides, how do I know you're not just telling pork pies? Matter of fact, all your posts are generalities, and whether you're saying stuff in ignorance, or just flat out lieing, I don't know. But equally, I don't believe you either.

I stand by what I said. End of story. ps: The party icon only proves I'm not interested in your opinion, crawdad.

 

[windude99]

Why are you all still arguing about this?

 

[SIW2]

Why are you all still arguing about this?

Maybe they enjoy it.

 

[windude99]

Why are you all still arguing about this?

Maybe they enjoy it.

 

[mart4494]

Why are you all still arguing about this?

Maybe they enjoy it.

Think some folks need to chill out over this. Reminds me on another thread about the virtues of Win8 over Win7 and vice versa.

 

[SIW2]

Nine pages of Yes or No is dreary.

Might as well spice it up a bit.

 

[HippsieGypsie]

Nine pages of Yes or No is dreary.

Might as well spice it up a bit.

And spice it up you gentlemen did! Now I know everything about Defrag as I ever wanna' know!

 

[AC]

Every time I look at this thread, I see arguments about defragmenting SSD's. I swear, on a lot of other forums I'm on, you guys would be banned.

Any way, I don't defragment SSD's. Never have, probably never won't. Now, I am going to

 

[poppa bear]

Truce

Sorry gang! Been fun while it lasted. But I surrender and hang up my double edge claymore sword. Crawdad can have the last say if he wants to.

Off topic: On a lighter note.

Reminds me of the story of a group of school kids who had to share class stories with a moral to them. When the proverbial Johnny was asked to share he said he had an aunty Karen who was in the armed forces. While flying over enemy territory, the plane she was in was shot down.

She bailed out with a machine gun, a machete and a quart of whisky. On the way down she drank the whisky in case it smashed when she landed. Turned out she landed in a nest of 100 hostiles.

She mowed down about 65 of them with the machine gun, till she ran out of ammo. Then she machetted about another 20, until the blade on the machete broke; and finished off the last 15 off with her bare hands using karate.

The teacher was horrified at such a terrible story, and asked Johnny what the moral of the story could possibly be?

Johnny replied that his father had said the moral of the story was: "Don't mess with Aunty Karen when she's on the booze!"

 

[Mystere]

I understand how file systems work, and as a software developer for over 20 years

I'm sorry, I wasn't trying to talk down to you. I didn't know what your experience was, so I was just trying to make sure I made myself clear and understandable to techy, but not necessarily programmer people.

I agree (and specifically said) that I do not advocate defragging SSD's. Just that it's not "ridiculous" to do so, just not all that practical. I said (repeatedly) that moving blocks around on an SSD has no benefit and was pointless, and does increase wear.

The thing about SSD's though, is that they are still treated as if they are physical drives by the hardware, and as such they are subject to a much more complex set of protocols and latencies that a directly connected set of memory would not be subject to (although it will be interesting to see SSD's that work over something like Thunderbolt, which is basically external direct PCI-e). So certain operations that would be negligible in direct memory access could be a lot more time consuming over SATA due to latency.

Yes, I agree that this is largely speculation based on theoretical usages, and I have not performed benchmarks to verify.

 

[Coke Robot]

Ok, seriously right now, why are we discussing defragging SSDs and keep dragging on a conversation that is turning into insults?

 

[HippsieGypsie]

All I know through experience is that I think I ruined a thumb drive by defragging it. I would be gun shy if I had or obtained a HDD. It was nice listening to the argument. Learned a lot.

I'm sure you'll all remain friends.

 

[crawfish]

I understand how file systems work, and as a software developer for over 20 years

I'm sorry, I wasn't trying to talk down to you. I didn't know what your experience was, so I was just trying to make sure I made myself clear and understandable to techy, but not necessarily programmer people.

I agree (and specifically said) that I do not advocate defragging SSD's. Just that it's not "ridiculous" to do so, just not all that practical. I said (repeatedly) that moving blocks around on an SSD has no benefit and was pointless, and does increase wear.

The thing about SSD's though, is that they are still treated as if they are physical drives by the hardware, and as such they are subject to a much more complex set of protocols and latencies that a directly connected set of memory would not be subject to (although it will be interesting to see SSD's that work over something like Thunderbolt, which is basically external direct PCI-e). So certain operations that would be negligible in direct memory access could be a lot more time consuming over SATA due to latency.

Yes, I agree that this is largely speculation based on theoretical usages, and I have not performed benchmarks to verify.

I think we're largely on the same page. I guess the problem with openly speculating in a forum like this is that someone such as the one proponent of defragmenting SSDs in this thread will jump upon your musings and turn them into "facts" (his word, repeatedly used) that support his wild claims, claims that have no credible, real support that I'm aware of, and numerous reasons to disbelieve.

 

[AC]

Right about now, I want to downgrade. To Windows 2K. Oh, how I love the Classic theme. Thank you Windows 7!

 

[pparks1]

I don't see a reason to spend money on this personally. It's got a few neat features, but certainly nothing that I cannot live without. Even at $40, it's money.

 

[AC]

And watch OS X start selling multiple editions. At the same prices of Seven.

 

[poppa bear]

Final thoughts on SSD

In fairness and the scientific principle of seeking facts, today I spoke to the professor who is head of the Dept of Eelctronics and Computer Science at Curtin University, West Australia. I asked if defragging SSDs is good, bad or indiferent. This is his answer:

No you definitely should not defrag SSDs for two reasons.

1. SSDs have a finite number of writes & reads, so defrags reduce the life span of the SSD.

2. The memory is differently oragnized so there is no need for it.

Defragging per se would have no detrimental effects, and in some cases could marginally increase the performance, but it is so small a gain it is not justified in respect of reason #1 above.

While I acknowledge his superior academic background, I am still not totally convinced, because his premise is primarily based on conserving the write/read life of an SSD. And this is definitely an open area of investigation, as shown by this typical blog by Max Schireson entitled: "Debunking SSD lifespan and random write performance concerns." It is found via this link HERE. In part he says:

Durability: The short answer is I wouldn’t worry about it.

For applications which are heavy on random writes, you’re OK (meaning a life span of over 5 years) up to about 25 million writes per day per drive, which is nearly double the IO capacity of the fastest hard disk drives.

For sequential write heavy applications (which benefit far less from SSDs), you’re OK (same 5 year life cycle) assuming the application re-writes each block on average no more than once per half hour; the smallest size of the latest fast HDD’s can barely manage this (take a Seagate Cheetah 15k.7 at 300GB for example, which has a claimed sustained write throughput of 171 MBps), and it gets harder as disks get larger.

And so it goes on. The point is, if durability is not an issue, there may some environments in which defragging may be of benefit. Note the reference to sequential writing and relate that to ability of hardware logical flip-flop IC chips to read binary codes once the data is collated, which is exactly what defrag does.

In summary:

• I am of the opinion that durability is not an issue.

• I strongly suspect there are certain environemnts in which defrag would help ... especially if the right type of defrag application is used. And this warrants further investigation.

Anyhoo, as Cokebottle and a few others have said, this is a thread related to Win8, so if anyone wants to pursue it further, should be done in an independent thread dedicated to SSDs. Cheers PB