PangingJr

Jeremy of Many,

in my opinion and experience, i always chose the software programs that i want to use from my own experiences with the software.
i do read just a few things that i need to know about the software products that i'm about to try it, because i don't believe a lot about anyone says about it, i just need to try it, and that is the only thing that help me decided which one is the best for me.

have you actually tried both defragmentation software products?
i suspected that you have, but have you?
i mean, can you determine by yourself which one is best for your needs?

don't get me wrong, discussion is good, but unfortunately, they are not always help you to decide what is best for you. it's always need your own experiences.

Jeremy of Many

Yes, I used PD for a year or more. I've been using DK for several months now. I go by my experiences as well, but I'm fed up with users of other forums saying one liners like "<Insert name here> rules! Very fast! Previous defragger ruined my harddrive, never go with it!"
Regarding defragmenters in particular, I like to get as much proper information as I can, especially since reps from both companies are involved in this thread.
Quite frankly, it's not my needs I'm concerned about, it's my harddrives'.

__________________
[b]Antec P180 | OCZ PowerStream 520W | DFI LanParty NF4 Ultra-D | AMD X2 3800+ @ 2.4 GHz | 160GBx2 + 500GB Seagate SATA2 NCQ 8MB & 32MB | 2GB OCZ PC3200 DC (2.5-3-2-5) | GeForce 6800GS 256MB | Sound Blaster Live! 5.1 | Razer Copperhead Laser Mouse | Standard M$ 101/102 Keyboard | Windows XP Pro SP2

PangingJr

okay.
please discuss away.

mmaterie

Hi Jeremy,

(I wrote this really long post to answer your first question before I saw your most recent really long one. I'll try to fit answers directly into what I had already written. Some of what is here may be redundant from your post). If I missed anything please let me know.

BTW: I really appreciate your investigation on this!

---

I should clarify that I’m the "product marketing" guy at Diskeeper. That means I head up the direction of the products, and what new technologies or features go into them. However, I do know a good bit about the file systems. Over the years, I’ve traveled to numerous Microsoft campuses in the US and Canada to train their Premium Support Services team on Diskeeper and file system performance. I actually started at Diskeeper Corp in a technical role, but moved into this position to manage the development of Diskeeper 9.0.

That said, I will spare you and everyone else in this thread a sales pitch. I’m here to help you better understand Diskeeper – that’s all. I hope that comes across in my response.

As anyone who followed that link may have read in the Diskeeper blogs, I don’t advocate or believe in head-to-head comparisons by a proprietary vendor. As I stated, there is obvious bias and a strong likelihood for incorrect information presented about the competitor product either out of lack of technical expertise with the product or intentional propaganda. I can't think that data would help users make proper decisions?

IT Professionals who spend company money want more than promises, they want results. I think that is exactly what the readers in this and other online tech forums demand from their hardware/software vendors. On that note, Diskeeper publishes papers that let you reproduce, typically with independent tools, that the product does what it says. For example the Benchmarking I-FAAST paper in our Knowledge Center details how you can verify it actually does improve file access. As I mentioned in the blog entry, I-FAAST is designed to also speed up new file writes. By design, you can get faster new file writes due to disk physics than with any other approach I have seen – that is of course just my experience of other methods. However, there is only so much influence a defragmenter can have on file system behavior, which is why we don’t heavily promote that or base a technology on it.

As for Windows XP's boot optimization, that was a technology co-developed by Diskeeper and Microsoft. MS built the prefetch system and the filter that determines the files and their order and Diskeeper wrote the code to sequence the files on the disk (for the built in defragmenter).

When I buy a product and a vendor tells me "we do this or that" I ask more than just "why", especially when I'm not an expert on the subject (like the nuances between Plasma and LCD). I make them prove to me that doing so actually makes a quantifiable difference. The reason they give may sound cool but be irrelevant for the purposes of the product. I see a lot of people doing this already (i.e. digging for truth) on this forum, but you have likely all seen people on some thread just buy into something on face value; hook, line and sinker.

If you visit the Diskeeper website you won’t see us talking about other products. I don’t mean to sound conceited, but we simply don’t need to – I cover more in one of my blogs on that. That also means we don’t infer things about other products. Inference is assumptive and contentious (my personal opinion). What if one politician said in a campaign "Someone in the government recently told me they thought that my incumbent opponent may possibly be fondling one of his employees inappropriately?" It either happened or it didn’t. If it did, prove it. It’s a marketing tactic (use of wishy-washy wording) used to encourage FUD (fear, uncertainty and doubt) and "unsell" the competitor. Unfortunately it works.

Do you need to "wonder" if Diskeeper did its job? The answer is simply, no. Does Diskeeper 2007 Automatic Defragmentation afford less control? The answer is again, no. Does I-FAAST "shuffle" files around back and forth, nope. Does placing files sequentially on the disk based on frequency of usage (not a rote file attribute) improve performance, abso-freakin-lutely! And, we actually prove it. Can you scientifically and accurately gauge disk performance without having to guess? Yes. It wasn’t easy, but our developers did it and we have the Intellectual Property (IP) protected. And, for the record it works better in production environments than in a lab because it dynamically adjusts to changes. If you used resources inappropriately to sequence files, sure, it could easily result in negative returns; that's true. Good thing then that Diskeeper does it intelligently and now invisibly. Does Diskeeper consolidate free space - yes, and we have IP protection (since 1999) on that as well. Has it improved over time - yes (just like the rest of the product), and it is an integral part of Diskeeper 2007.

Please note that Diskeeper is actually gauging file usage frequency. Reading last modified date or last access date file attributes does not substantiate sufficient knowledge about usage to justify saying that an ongoing action is based on "frequency of usage". The I-FAAST system is far more advanced. It is actually learning about your system and how much of your PC's time is spent reading existing files, modifying them, or creating new ones. All that info let's it speed up file access. I'm not saying another defragmenter needs to go to those lenghts for it's file strategy purposes, but for I-FAAST to deliver on its promise, all of that is vital.

Many of the above comments are key points of development effort in the new version. Without InvisiTasking you could not run in real time. We install Diskeeper in an automatic mode because we are 100% confident in the technology, we know it works. Why schedule when you don’t need too? The new graphical control panel for Diskeeper 2007 in the dashboard offers greater flexibility that ever before. We built the ability to turn if off at certain times for those who are very sensitive to the thought of real time defrag (we understand it is a new concept for many and it can take some getting used too).

There is a falsely propagated myth that Diskeeper does not do free space consolidation, or de-emphasizes it. That is simply untrue. What we don't do is over-hype it's value. It is a popular topic and I did cover it in good depth on the Diskeeper blog. I hate to keep referring readers off this site, but it is a whole separate and very lengthy topic.

That leads me to another thing any software/technology buyer should consider; white papers that originate from, or are funded by, a vendor should always be suspect. I don't mean to sound like your momma (my apologies to anyone reading this if I've come across that way), but you know that many people believe everything they read or see. Don't get me wrong, I’m not saying all vendor papers are bogus, just a heads-up to be careful. In most cases, third parties are very ethical and would never compromise their ethics so most analysts can be trusted.

(added note: I see that you just brought this up to Jeremy)

Let’s take a hypothetical example. Hire a very intelligent guru/analyst or solicit one that works for your company. Have him detect that a certain operation of a competitor’s product, that if run in some non-standard/reverse manner, makes it return a bad result. Point to an obscure, non official, comment that supposedly justifies the operation of that competitive product in the manner undertaken in a test. And then present the results, purporting that the test is a true apples-to-apples comparison.

An example might be a comparison between backup software vendors. What if a head to head test involved a speed comparison? If, for one product, all the bells and whistles were turned on and on the other product nothing but the basic "back up the data" was initiated, it wouldn’t be a fair test would it?

Unfortunately that kind of testing happens more often than it should.
I’ll end off with the statement that Best is subjective.

If our advertising team ever says "best" or "leading" shame on them, unless they quantify it with facts. Yes, we may have used Number One, but that's substantiated with sales figures (90%+ market share, as measured by National Purchase Diary – the independant group that tracks sales in the software "channel"). That slogan was used for those unfamiliar with Diskeeper’s market share.

While sales are a pretty good indicator, it doesn’t mean a product is the best. I think we can all agree to that. That’s why I feel it is our [Diskeeper Corp] responsibility to continually evolve the product and improve it technically. Symantec (Norton Speed Disk) was the big kid on the block in Windows defrag in the 90s. Diskeeper had a better product and supplanted them, as shown by sales. Granted Symantec, to their credit, focused on Security as they could make a lot more money in that market, and competing with our defragmenter wasn’t worth their investment.

Perhaps the best thing about most commercial defragmenters on the market is they offer a 30-day trial period so you can test drive the software first. Diskeeper offers an unconditional 30-day money back guarantee as well. Other vendors may do the same.

Best regards,
Michael Materie

PS: I also write all our technical papers such as on virtualization, system reliability, operating system architecture (how it relates to our products), file systems, etc…so if you ever have a question, please stop by the diskeeperblog and ask me. It is a 99% "marketing-free" zone.

Jeremy of Many

My pleasure.

As I understand it monitors the routine of the file system activity and adapts to any changes made as well. Does it work in conjunction with Free Space Consolidation?

The next two quotes are key in my investigation.

(From Michael's blog post "Comparing I-FAAST")
So frequency of use and modification date are not the exact same, thus indicating the difference between DK's and PD's defrag methods. Is either one better than the other?

More of what I was looking for. Would you say, for anyone to get to the bottom of "Which defragmenter should I use?", could and should rely primarily on the blog posts on both www.diskeeperblog.com and http://perfectdiskblog.typepad.com/perfectdisk_blog ?

So you're saying I-FAAST is better than SmartPlacement?

I totally agree with you on scheduling vs. automation regarding PD and DK. However, I want to acknowledge my realization that it is my preference and in no way whatsoever is that influencing my priority or determination to gather as much proper info as possible regarding either PD or DK in this thread. If I decide in the end, based on the info that Michael and Greg give me, that PerfectDisk is better for my harddrive, then I will return to the scheduling functions of PD.
How considerate of you.

That's fine. It's right here
The statement that stands out for me the most is:
So if I have 5 files, at 5 fragments a piece (as if I were to defrag and reboot have those user profile .LOGs and .DATs changed again), I don't really have much of a problem yet...
However, for the majority of users out there, a problem does exist.

Yes, I have several friends who have seperate partitions for C:\Documents and Settings, C:\Program Files and C:\WINDOWS. I, myself, haven't done this nor do I have the intention of doing so anytime soon. Maybe later. It makes sense that by doing so, it would reduce fragmentation (by seperating dynamic data from static data. I do, however, have two harddrives. One contains Windows and all installed programs and another contains all stored data (archives, games, movies, music, software, various documents).

In most cases, yes. However, regarding defragmentation software, one has to be healthier for the drive than others. Unless your answer to the next question is "Yes".

Are you saying that neither DK or PD are better than the other when it comes to defragmentation as a whole?[/QUOTE]

__________________
[b]Antec P180 | OCZ PowerStream 520W | DFI LanParty NF4 Ultra-D | AMD X2 3800+ @ 2.4 GHz | 160GBx2 + 500GB Seagate SATA2 NCQ 8MB & 32MB | 2GB OCZ PC3200 DC (2.5-3-2-5) | GeForce 6800GS 256MB | Sound Blaster Live! 5.1 | Razer Copperhead Laser Mouse | Standard M$ 101/102 Keyboard | Windows XP Pro SP2

mmaterie

Hi Jeremy,

I feel it's not my place (given I represent a vendor) to say one is better than the other or dish out sales pitches in a technical forum. That is for you and other readers to discuss. I'm here to help explain how our products work and why they do what they do, and I think Greg is doing the same. My purpose is not to attack Greg or Raxco's product. Raxco makes good products and Greg is a smart guy who has repeatedly earned the MS MVP award, which proves his desire to help by sharing his knowledge.

I'll piece together answers to your latest questions:

-As for I-FAAST being better than SmartPlacement, that is for you to decide. I think the question is really what is better for you; which depends on what is important for you. They are different technologies with different purposes. How the technologies gather the data is relevant only in so much as how they need to apply the data they get. If Diskeeper's I-FAAST were to rely on Last Modified Date, it would limit it's effectiveness dramatically. The only suggestion I have, is that when you are evaluating what is better, that you get pick the best product based on real results, not based on what vendor has the most attractive sales pitch.

-I-FAAST does implement free space consolidation as part of its overall performance strategy.

-Keep in mind that when you use a third party file system performance utility to sequence and/or place files, (especially I-FAAST) you'll get better performance that partitioning by data type. I still recommend certain partitioning strategies based on disk spindles - such as placing the paging file on a separate spindle than the boot partition (i.e. \Windows).

-If you prefer scheduling over automation, that is your prerogative. As I noted before, the ability to turn it off, makes it relatively similar to scheduling in that you can decide to run it only when you allow it. I just spoke with a major bank today that is rolling out DK 2007 to all their workstations. They are turning off DK during evening batch jobs, not because it's not invisible, but simply because it's company policy to never run anything while the batch jobs run. It's not my business to tell them that's wrong. What's right for them is what's right for them.

-The blog sites are the vendor's way of providing data in a new format. I'm a big fan of Microsoft's blogs - there is an unbelievable wealth of info on their blogs. Much better, in my opinion, than on their website. As a vendor of any product there are company rules and procedures for promoting and selling and narrow channels for getting that information out to customers. In most cases they are conservative and relatively terse. Company blogs give individuals the ability to expound on whatever they want to talk about in a less formal manner and without a great deal of company effort (often official company docs are heavily scrutinized and go through a long approval process before publishing). I agree with your comment that most company blogs are good sites to get info. I try to make the Diskeeper blog a purely technical site. We may slip in the occasional PR announcement, but you can always skip over those if you don't care to read them. I've have seen blogs that only serve as a marketing extension and don't really provide value.

-On your question about number of fragments, it's all relative. Humans, per studies from research fields like HCI - Human Computer Interaction, are less sensitive to slows than applications - at least from a performance standpoint. Us humans can usually live with a few seconds delay before we complain. That's not to say a computer can't, but they move inexorably faster, so holding up one application a few seconds is far more pronounced from a performance standpoint. In other words, applications tend to be more sensitive. Your applications will definitely notice 2 seconds (such as video editing for example). Imagine if you bought a DVD on Amazon and the purchase transaction was delayed by 2 seconds - you'd live with it. However, if you're an Amazon exec and every transaction was slowed by two seconds, it means a whole helluva lot less money for you. The point is that 5 files in 5 fragments each is not going to make your PC noticeably faster - at least not by you. Yes it will get worse, and in almost all cases be noticeable by the user. Fragmentation will start to affect your computer's ability to produce well before the human interactive experience is impacted.

As the product manager I think Diskeeper is close to perfect, but I also know there are things we can improve or newly invent (we've already started on the next version, and I have another whole version nearly planned after that one). What I always want to hear is what else we can do to make it better (wishlist@diskeeper.com). I think that holds true for almost every manufacturer/service provider in any industry. Companies that don't listen to their customers are doomed to fail. Raxco products have improved over time as well.

-Now, can you test what product uses the least overhead to keep the system tip top - yes. Can you determine which product incurs the least (or no) effective overhead - yes. Can you test which product speeds up performance the most - yes. Will one product generally do better than another - yes.

I understand that most consumers don't have the time, or ability, or even the desire to go through all that testing, but I don't really think they need to. I just don't see grandma running PCMark or Perfmon on two live running identically imaged volumes to test I/O overhead and file read time benchmark scores! System Administrator's, on the other hand, do this kind of thing.

Diskeeper is a great solution and I believe its results and reports go a long way to proving that.

-Michael Materie

Chaos

once again I want to thank Raxco and Diskeeper staff for making this thread as informative and interesting as possible without attacking each others product line.

__________________

rms13

To Jeremy:

I think that we are asking here the wrong question. It should be - not which one is better, but first - do we need to pay $40 or $100 in the first place.
$100 is a half of what Windows XP Home costs, so it is better to think twice before purchase. Of course, I am talking about home user - on the server, need for automatic defrag is usually undeniable.

No doubt, commercial people of both companies are working hard, but they are not telling the whole truth, more like creating an illusion that we can knowingly make our choice. But, how can you compare, and have anybody actually done that?
This business is somewhat like that with oil additives. People want to believe that some product will do magic, and will better pay money for such 'miracle product' than change oil as appropriate.

Have a look at the examples to prove how much defragmenters will increase perfomance. I have no idea where they have obtained such disks, but they are not from the real computer. I have serviced computers for years, but haven't seen anything similar. Even on NT computer that was working for at least 6 years without any defragmentation 24/7, disk was not that bad.
But you should compare to normal disk, not over-filled, defragmented at least weekly with Windows defragmenter. How big the effect will be then, really $100 worth?

Here are some results.
Test computer is Toshiba P4 laptop, with 1G memory and 60G 7K60 drive. It is partitioned in 34(C drive) and 22 GB partitions, C partition 40% free, data storage partition 5% free.

First, Diskeeper2007-ProPremier latest trial version was installed, and after 1 week of operation (at least 10hours/day, IFAAST enabled), measurements were done - Windows startup till login screen; Lotus Notes till login screen and till messages are displayed; Photoshop startup; Autocad startup. Lotus notes was used daily, Photoshop and Autocad - just for test.
Then, Diskeeper was uninstalled and replaced by PerfectDisk8 latest trial version, smart placement defragmentation(with startup optimization set to 'let PD manage) and the same measurements performed.
Then, PD uninstalled, drive defragmented with Windows defrag(it significantly restructures disk - layout.ini file placement is completely different in PD and Win approach) and again the same measurements. All given times are average of 3 measurements in seconds.

Diskeeper Startup - 33.0; Lotus - 6.4 + 7.7; Photoshop - 18.3; ACAD - 27.3
PerfectDisk Startup - 33.6; Lotus - 7.2 + 8.4; Photoshop - 17.9; ACAD - 28.6
Windows defrag Startup - 31.2; Lotus - 6.6 + 7.3; Photoshop - 18.2; ACAD - 26.8

Now, have I got it wrong, did I measure not correctly, or I have not measured the right thing - but where are those $100 worth improvements ?

And then, if we want to see what really slows down the computer, do a simple experiment. Create something like 50 000 small files. It can be done with script like this (in command window) - FOR /L %f in (1,1,50000) do echo abc > %f . It will create 50 000 files with text abc in current directory. Then select 10 000 files, right click and enjoy. Note that all those files are stored entirely inside the MFT, so no defrag can do anything about it. And, by the one, you will have discovered a free substitute for Frag Shield. You can delete all those files(it will take time), but MFT size will stay increased (~1K for a file), as it cannot shrink.


To Greg/Raxco Software :

As you see, your boot optimization does not work. It does make much sense when described on paper, but in the reality Windows defrag places those files very differently and wins(I have measured it on other computers, and results were even more impressive in favour of Win defrag).
I have raised this question with your support, but the best answer was that I am free not to use the feature! Obviously, they were in belief that I have paid for the software already. Maybe you have any better comment?

Smart placement - I have tested with some dll's in system32 folder, which are not listed in layout.ini file. Simple cnanging last modified date and next defrag pass moves the file to the end of drive. You may call it smart, but it does not seem to me that way. Just wondering about patent system. Maybe it is the reason you haven't included the feature to locate file on disk. Your tech support redirected me to some article which says it is not done for sake of resource economy. I am running DiskView and wondering, does it really use that much resources.

To Michael/Diskeeper :

As you see, I haven't noticed any measurable perfomance increase from your $100 program. Could you please explain, where I am wrong?

What I have noticed though, is that if run in manual mode on data disk (5% free), it performs like simple multi pass defrag - every time it is run, it is trying to move some data around, works for some time, but fails without doing the job. Looks not better than Windows built-in defrag.

Manual Analysis Job Report displays information that is simply not true - something like 'Warning! The MFT usage is currently 90 percent of total MFT size, which indicates it is likely the MFT will become fragmented.'
Why would it, if there is at least 500% of reserved space? Is this statement intended to scare somebody who does not understand how MFT thing works, and to help sell the air bubble Frag Shield?

Finally, and most importantly, the idea behind IFAAST is really great for server.
But, what about home computer or even workstation? What makes you think that what I do most frequently, is the most important thing I have the computer for? It could be a computer on the Shuttle, where I maybe could play games for 2 weeks - but I still need it for landing.

If you really wanted to make an 'intelligent' defrag for home workstation, and it applies to PerfectDisk too, the best would be to speed up the system, not applications or data file access times. But that does not require any real time learning - that is, if you know how Windows works.
Or, if you really wanted to justify $100, it could be IFAAST with 'learning mode', where information woud be gathered about most important activities and then applied to manual defrag passes - without anything automatic running in the background (at least, as an option). You are not going to say that you can monitor file system without using any resources at all - to the people who have disabled most of Windows services and gained a measurable improvement?

Of course, current automatic mode could be left for those who want to believe in magic - would that be most of your customers?

Thanks.

mmaterie

Part 1:

I agree that everyone needs to evaluate these products. Every general purpose file system in a commercially available OS generates file fragments over time - that is fact. I'm assuming the people reading this thread are familiar with that, but if anyone wants proof, google it and you'll find plenty of reputable sources.

I don't think any major defragmentation vendor is pushing anyone to buy out of fear. If some communication comes across that way, let the vendor know - marketing people make mistakes. I can't speak for all vendors of this type of software, so some may not be ethical. One unethical vendor does not mean that they all are.

The primary difference between a server and a workstation is that a server hosts more data, users, etc. Every defrag vendor I've seen discuss fragmentation states quite openly that fragmentation is proportional to the amount of file modifications, writes, and deletes. If the computer does nothing but read files on the disk all day long, apart from background system activity, you're not going to see much fragmentation - and it's highly unlikely to become a performance issue.

If you use your PC for internet browsing and web based email, you can probably go a very long while before a defragmentation job makes a performance difference for you.

Due to my industry affiliation, I receive a great deal of free commercial software. I have a half dozen antivirus and antispyware programs sitting in unopened boxes. I have free licenses to install them on my home PC's but I do not. It's not that they aren't good products - they are. I don't install them because I don't need them. I know to avoid certain places on the internet and to not open attachments. I have a background in security (I used to work in a technical role in that industry). For another person, that may not be the case, and anti-malware may be a crucial part of keeping their PC's operational and performing well.

If you try a piece of software and you can't perceive any benefit from the software, DON'T BUY IT.

It's unfortunate that you feel there is withheld information here. I already stated earlier in the thread I understand that home users are not going to have the time or wherewithal to undergo scientific analysis. That's why the media (who speak freely), and the idea of freedom of speech are so infinitely vital to protect consumers/citizens. In the PC world, there are numerous media experts and IT analysts out there to help buyers/companies make the right decisions.

Exactly how is this business akin to that of selling "oil additives". Is that entirely due to your evaluations? So that I'm not misunderstanding you, what exactly are you insinuating? Is everyone else who recommends defragmentation a snake-oil dealer or friends with one? So Mark Russinovich, Paul Thurrott, Jim Allchin, Walter Mossberg, NTFS file system developers, and hundreds more are all on some defrag vendor's payroll or just hawking thier own mystery oil? Is any vendor telling people not to "change their oil" per your analogy.

I understand your are segmenting the consumer from the business machine/server, but that isn't the dividing line. It simply comes down to how much a computer has to do with respect to disk I/O dynamism; company computer or home PC.

I understand your speculation and understand personal experience drives your beliefs. I can't argue that - cause I'm the same way.

Keep in mind that major defrag manufacturers make 90% of their income selling to businesses. Diskeeper marketing (with exception to technically simplified emails sent to home users who download trialware) is 100% geared to IT professionals.

Apparently I need to re-iterate that I'm not in this forum to sell 2 or 3 licenses to "gullible victims". And, I'm not here pushing people or scaring them to go buy. I'm here to help answer questions that users have about Diskeeper. That's it!

I feel like singing a Mel Brooks tune. Maybe something from History of the World Part 1. Something with dancing nuns should do...

mmaterie

Part 2:

I think you may have misunderstood some of the purposes of the products and how they work. A defragmented file is a defragmented file, no matter what product you use to get it there. The differences are whether one product can get there, and what resources/impact it takes to do so. In a corporate market, other factors like automaticity and management capabilities also weigh in more heavily.

For a scientific test you need to always begin with an identical starting platform, and you can't sequentially run one product after another. I'm assuming the tests were timed programmatically (i.e. not by stop watch)?

Are 3 test runs sufficient to come to a conclusion? If you took statistics and probability courses you may recall the variable for error on only 3 tests? http://en.wikipedia.org/wiki/Margin_of_error.

Was caching addressed to ensure the reads were direct from disk?

I also don't believe Raxco makes claims that file access is sped up beyond defrag with their features (they claim to slow re-fragmentation), but I'll leave the detailed explanations to them.

Keep in mind I-FAAST is an ongoing feature - not a once and done manual job. As your disk fills and the layout files scatter (it will happen - just look inside the file to see what is in there) Windows will end up, over time, pushing the file further back on your volume as it seeks to find a single large chunk of free space to place the entire group into. That is due to NTFS file allocation algorithms which will fragment the free space as it writes newly modified files nearby the original. As that happens, watch your boot up get slower and slower. Maybe the layout was ideal that day? Also, what did I-FAAST infrom you it would provide performance wise? Can you confirm that the files/processes you tested are now considered by I-FAAST to be frequently used and I-FAAST has actually processes them?

If you run I-FAAST prior to the built-in, I-FAAST has already placed the boot files. Why would the built-in move them?

If you don't reboot, log out, disable all services, and never use the disk for anything (store new files on it, etc...), then those numbers will stay the same.

Yes, creating files causes significant I/O overhead. Is that your point?

That script, which has been available at numerous system tweak sites for years, is not unlike what Frag Shield does.

Sounds like a rhetorical question to me, but I'll bite anyways.

In addition to all the questions on the testing that I've already asked, was anything related to the Photoshop, Autocad or Lotus fragmented initially? If not, defragmentation won't improve performance. Where were they were located to begin with? What tools did you use to measure (bootvis)?

If you are an IT Professional (or scientist, or mathematician), you know that results are only valid by following a scientific method and minimizing/eliminating the error margin. Following and documenting the method itself is just as important as the results.

If you review the Benchmarking I-FAAST paper, you'll see several pages of test methodology. When I write a paper I ensure anyone else can repro the results. If someone else can't repro test results, you do not have a case - you have an anomaly.

I'm not invalidating your testing, I'm just pointing out that your methodology, in detail, is absent from your results.

I agree the warning is dramatic, I can get this corrected in a future release.

I'm not sure where you get 500% from? Your upset with Frag Shield is duly noted. But I'm not sure why? According to your logic the MFT would never fragment? I don't think I need to go into detail on that aspect as it was covered earlier in this thread, do I?

For the record, Diskeeper is not designed as a manual defragmenter. Very little improvements are made in that area. As I noted, Diskeeper is an enterprise product focused on running automatically and silently. Multi-pass is a strategy for minimizing resource usage (specifically I/O overhead) - it was never built for manual defragmentation begin with. If you want a manual defragmenter, I'd suggest you look at other products.

I-FAAST speeds up file access above simply defragmenting. It's also continually learning and updating its information so as the usage changes, so will the file sequencing. And, if a file is ideally situated already, it keeps it from moving elsewhere where access speed would be reduced. In v2007, we also added the ability to hand-select files should situations as you describe be the case.

You're entitled to your opinion. Given that the disk is typically the slowest component in a modern PC, optimizing that component returns results. I agree that there are many drivers/services that many people won't need and the Windows Service Host is getting more and more bloated, but we need to remember that Windows is a general purpose OS. It is kind of a one-size-fits-all system.

There are many forums and guides that anyone can follow that educate on how to tweak the system, as well as products on the market that automate it.

I'm not a world-class expert on Windows and file systems, though I am very proficient in their design, architecture and operation, and have written several technical papers that discuss these subjects already. I study up on it, and personally explore the system with various tools, so I can be better at my job working with our developers - who are world class experts on file systems and operating system internals.

I'm not sure why you are goading me here? InvisiTasking relies on default Windows system monitors to operate. I've noted this on other forums as well.

These aren't the droids you're looking for...

As I noted, most of our customers are IT Professionals. Are they fooled by slight of hand? I seriously doubt it, but I don't know because disk defrag is legitimate. I know they don't by RAM-defragmenters if that is where you are headed. IT Professional's test the product for weeks/months and benchmark performance on cross sections of desktops and laptops. They use various tools like PerfMon (looking at I/O counters) and benchmarking software to justify a purchase.

I also don't think a great many users active on technical forums are as gullible as you suggest either. Are there people, as Jeremy pointed out, that simply conclude "it rocks!" because of a pretty UI, of course.

Anyways... If anyone wants to ask me about how Diskeeper works, I'm always available to answer questions. You can contact me at the company blog, anytime. I'm restricted for time the next few weeks, so if I do return here to answer some questions, please understand I can't keep up these novels. Thanks for bringing up some good questions and considerations.

-Michael Materie

rms13

I am not arguing with this. Quite contrary, I am using DK on servers and recommending it to almost anyone as the best defrag for server. What I am saying though, is that I haven't seen any proof that it could show any measurable difference on the properly maintained home computer - that is, on which Windows built-in defrag is running at reasonable intervals. That is true, but in essence the main difference between server and workstation is more like difference between transport flow in the city, and one individual car. You can consider the first measurable, predictable in most ways. And it is done so when creating transport flow control systems. While, other-single car, may be predictable too in minds of creators of, for example, automatic intelligent gearboxes, but in fact is not. This is a mistake that serious car manufacturers have done(like even BMW), and defrag software vendors are doing too. The computer in test was not used for web browsing. Instead, it was intensively used whole week for preparing a project - that is why 10hours/day. The most used programs were Bentley micro station, then Visio for preparing diagrams, Acrobat for creating pdf's. The only thing - files were stored on data partition and ready files transferred to server.

And, as using Micro Station was kinda unusual, I have specially tested Autocad startup times instead, as it is in general much more important application, along with Photoshop. That is a real life situation, you know, not some kind of artificial test drive.

<br />
Exactly, but as I said already, there is no way average user could really test it. Instead, user may(and does) feel better, because this is a general mood(created by vendors) out there. Similar as so many are buying different cleaners that simply find *.bak files, or registry tools that do very primitive job, but makes people happy.

Probably not, or maybe yes, but at least Mark Russinovich has created something like contig, which if paired with power defrag, is FREE and quite adequate for every home user needs.
Unfortunately, I haven't seen any vendor advertising similar free solutions.

I see server exactly as a dividing line. Completely different concept, and different approach needed.
Thank you for that.
I hope you will help to understand how your product is $100 better than properly used Windows built-in defrag, combined with good general computer practices, then.

rms13

I haven't misunderstood anything.
You are advertising that you are accelerating file operations. I am showing that in the real world it will not give any significant improvement, except some satisfaction from newest latest program working on your computer.

I completely agree with you on corporate market, and I am not arguing with you there. I am talking about home user, who is made to believe that those same corporate benefits will apply to him too.

It is not a scientific test.
I am just showing that, on the real properly maintained system, your program does not show any noticeable gain, as contrary to your specially conditioned test systems.

And yes, I can run products sequencially in this case.
I have given one week for your product to improve perfomance on the computer, previously regularly defragmented with built in defrag. If you are interested in results before DK, here they are:
Startup - 31.3; Lotus - 5.9 + 7.2; Photoshop 18.4; Autocad 26.6

If it didn't do anything in 1 week, that is exactly what I wanted to show.
For PerfectDisk, I just wanted to show that famous boot optimization, which is different from MS one, does not show advertised efffect. Everything else is so straightforward there, that nothing to test, in fact.

And, tests were timed exactly as they should - by capturing the whole process, including stopwatch, on the tape and later calculating the result. I have done that before with bootvis, but that is yet another program running.

Are you seriously suggesting that 100 tests would reveal something unusual? I have done enough tests before, and, unfortunately haven't noticed that yet.
I can reveal that I am also a victim, and kinda looking for that miracle to happen, too. And I somewhat believe that defrag could be made that would really speed up system a bit. It just didn't happen to see one yet.
What caching? As you see, computer was switched off every time.

That is correct, even that claim is also not really serious, based just on last modified date. But they also claim to improve boot time, which does not really happen.

I think that MS have improved the boot thing recently. And, now they move layout files to some specific position evety time you run built in defrag. Even if PD relocated them, next built in defrag pass puts them back to MS suggested place, not some free space.

I didn't say built-in moved them after IFAAST, PD did. Obviously, differences between built-in and IFAAST are caused by other factors, not layout file placement(if it really works exactly the same as built-in).

Computer was switched off every time.

It is not only creating. The point is - any operation in folder with many files causes that overhead. It can be 5 minutes or more till you get to right click menu when selecting 10,000 files.
Yes. Could you please explain, what is the purpose of reserved space (other than to just occupy the best disk location), if you are increasing MFT itself?

As I said before, computer was properly maintained, which is the whole point. If I have to specially fragment files to see any effect of $100 program, then sorry - I didn't do that.

Sorry, could you please repeat a link to a paper which explains how to reproduce results on real computer, compared to defragmenting with built-in defrag?

Thanks.
Exactly, MFT will not fragment until ther is reserved zone adjacent to it. And that zone will not disappear until you do not abuse (i.e. overfill) the disk. 500% is that reserved space was about 5 time larger than MFT itself.

Fine, just tell me please, how can I verify that on real home computer (which is not intended for performing continuously repeted tasks, like server or corporate computer may be). And not compared to some specially trashed computer, but properly maintained one.

Thanks.

gshayes

Folks, sorry for the delay in responding. My day job keeps me busy and I don't hang out in the online forums like this constantly

Couple of things:

I had earlier mentioned Raxco's patent-pending Resource Saver technology. This is technology that has been around in PerfectDisk for a while. Raxco started the patent process a while back (folks may or may not know how long it takes to patent something) and shortly before the launch of PD8 received the "patent-pending" nod from the US Patent and Trade Office. This technology results in less disk I/O required in order to defragment a file and was one of the reasons that PD was the first to support multi-tb drives and today will run on even larger drives where most other defragmenters really struggle. From a workstation perspective, do you care? Probably not. From an enterprise perspective (production servers), this is really important.
-------------------------
I think that there are some mis-perceptions on how PD works in conjuction with Windows prefetching and layout.ini. PD can either be configured to manage the boot files (subset of layout.ini files - relating specifically to system boot), to manage all of the layout files (all of the layout.ini files), to let Windows manage these files, or to completely disable so that nobody is managing these files. With the default setting of having PD manage the boot files, what PD is doing is telling Windows to NOT manage these files - meaning that Windows is configured so that when the idleprocess tasks run, Windows no longer executes defrag.exe /B (defrag.exe is the command line interface to the built-in defragmenter and /B tells it to do a partial defrag of just the boot files). This is to eliminate any "thrashing" between having PD move the boot files to where it wants them to be and Windows moving the boot files where it decides to place them. Note that during a normal online defrag by the built-in defragmenter, the built-in defragmenter reads layout.ini and will NOT move the files - it knows to leave them where they are located on the drive. If PD is configured to let Windows manage these files, then PD will mark these files as excluded and NOT move them at all. If PD is configured to disable, then neither PD or the built-in defragmenter are going to do anything with these files. Note that if Windows is executing defrag.exe /B, the built-in defragmenter is going to try to "move" the boot files to the first place on the drive that will hold all of these files - which is typically toward the end of the drive as that is where typically the free space is located (assuming any sort of free space consolidation is occuring). Note that when performing the partial defrag of the boot files, the built-in defragmenter does NOT create contiguous free space to put these files into - it depends on that contiguous free already being there. For the vast majority of systems, having PD manage the boot files results in anywhere from a 10-30% improvement in boot speed. Will all systems see the high end of this? Nope. Will all systems see the low end of this? Nope. Are there cases where no improvement in boot speed is seen? Yep. Are there case where an improvement is made but people don't necessarily notice (ie boot speed goes from 30 seconds to 20 seconds)? Certainly - but you can't deny that an improvement wasn't made. Are there people that think that Windows can do a better job of managing the boot files? Yep. The nice thing is that with PD you have a choice.
-------------------------
Regarding benchmarking on a system by using your stop watch to measure boot speed and application launches. Some people see no improvement beyond what the built-in defragmenter provides. Others see a major improvement. It really depends on how fragmented the boot files and application files are. They are lightly fragmented (or not fragmented at all), then it is certainly possible for you to see no improvement beyond what the built-in defragmenter provides. But keep in mind that with these types of benchmarks, you are measuring read performance - not write performance. That is why I'm a big fan of independent benchmark tools that measure both read and write performance.
-------------------------
For some people, the built-in defragmenter does an adequate job - especially in the optimal environment of plenty of free space and no very large files (the built-in defragmenter requires contiguous free space the size of a file in order to defragment that file). However, you hear more complaints about the built-in defragmenter and it's many limitations than you hear people praising it
-------------------------

This has been an interesting conversation but I have to get back to my day job. The best advice that I can give to people is this. Just as you try 3rd party anti-virus/anti-spam/email clients/etc... don't feel shy about trying a 3rd party defragmenter. Only you can decide if the cost of the software is worth the benefit that you feel that you get from it. If you don't feel the benefits are worth the cost, then if you are happy with the built-in defragmenter then stick with it. As I mentioned earlier - it does an adequate job for some people.

- Greg/Raxco Software
Microsoft MVP - Windows File Systems

Disclaimer: I work for Raxco Software, the maker of PerfectDisk - a commercial defrag utility, as a systems engineer in the support department.

mike2h

ty for time & info! fwiw i notice an improvement in bootup with pd over win. not much but enough thatim very happy with your prog. still got to try dk though

__________________

Robert McClelland

Hi this good thread. I'd love 10 diskeeper pro premier so far I love it You do have do boot defrag too that does help. I'd used Ifaast all the time. IS MFT really that help? I did with my hard drive I ran it I don't notice that it hurt it or not.

__________________
"XTREME AUDIO IS CRAP!" "There no support for it. It's crapist card on face of the earth it should be throw into the sun"
http://www.speedtest.net/result/1949536385.png

PangingJr

i've seen some of info that was pass by my reading in the past about people were saying that they had problem with their software/firmware RAID configured drives when they used Diskeeper 10, so this is not just the info on this specific thread alone. however, i also used Intel software RAID drives at the time i tested the Diskeeper 10, i remember i tried everything on the software program, its Frag shield and all. however, there was no problem with the file system or anything that were on my RAID drives at that time, but since i only tried it for like a week, so i'm not sure whether or not there was exist a real problem with Diskeeper 10 and RAID drives.
there was one thing that i also remember that i had to do after i finished testing the Diskeeper 10 and that was to completely format my hard drives again and than to restore all the files and the OS partition itself back on to it from my backup. this because i was not sure whether or not i should leave all the file system that were in the partitions like that after i've played around with the Frag shield's, and since i was not going to use the Diskeeper 10 anymore and i was too lazy to find out any info at the time so i just did that.

ScythedBlade

In the end, I got so confused, so I did some researching. So, here's the thing that I believe everyone should refer to:

PerfectDisk deframents your drive so that you don't have to defragment it so much. It's because the algorithm it uses places the most used files on the inner tracks. Theoretically, most used files tend to update themselves. So they place new files after. Therefore, it really doesn't need to defragment itself that much.

Diskeeper, on the other hand, places more used files near the outer tracks. Therefore, you should see somewhat more of an improvement in accessing files of the recent one. However, it fragments extremely easily since the most recently files updates itself and may run outta space, making it into the inner tracks. This is usually why diskeeper needs to constantly defrag.

See how your hard drives work. If you have 25 GB on a friken 750 GB hard drive like I do, nothing really matters. However, if you have a computer that's really wasted (dual core C2D overclocked 70%) you might as well use diskeeper. It'll wear out your hard drive a lot faster than PerfectDisk will. If you want efficiency, perfectdisk. Performance, diskeeper. (it's like asking if you want that core 2 duo to be overclocked or speedstepped ... goddamnit, if it was only this easy with defragmenters)

Jeremy of Many

I guess the sad truth is that there really is no 100% "this is my spot for best performance" area for each file. I'm starting to think that each defragmenter just does the best it can given whatever algorithm it goes by. Diskeeper confuses me with all the percentages and statistics it comes up with when you open the Diskeeper.msc. Minimal access time and percentage, gibble gobble. I like that it defrags in "real-time" (aka whenever your system has idle resources for it to run). However, for those of you who do anything CPU intensive on a constant basis like SETI@Home or Folding@Home which runs in low priority 24/7, get ready for Diskeeper's InvisiTasking not working at all, ever. I've been doing the SETI@Home for a few weeks now and my C: drive (which is 150 GBs and has 7.21 GBs used space was at a 4% fragmentation level. The excess fragments were massive! the moment I stopped SETI@Home, the fragments went down to a mere 8.
But that's just defragmenting, not file placement (which is where PerfectDisk would kick in) and what Diskeeper says "this file could be placed here to minimize access time). it makes you want to use both PerfectDisk and Diskeeper, which would be either harmful or redundant, or both.
That brings me to my next point: Diskeeper's InvisiTasking is essentially like transparent scheduling. It is set to always run, wait for idle resources to operate. It's automated. PerfectDisk's AutoPilot is essentially the same thing, it just requires one extra step from the end-user.
In my opinion, the difference between these two applications is becoming smaller and smaller as I look into this.
Their marketing is crap, just ignore it. get down to the nitty gritty to understand their functionality completely, but then back up a few steps and look at this in simpler terms.
They both defragment. One places files sequentially for best access time based on a file frequency of access algorithm, and the other does relatively the same thing based on a file modification date algorithm. The difference between those two algorithms is what... a millisecond of access time for some files?
If that's true, then does it really make the slightest bit of difference whether you use Diskeeper or PerfectDisk, or UltimateDefrag or O&O Defrag or Vopt?

__________________
[b]Antec P180 | OCZ PowerStream 520W | DFI LanParty NF4 Ultra-D | AMD X2 3800+ @ 2.4 GHz | 160GBx2 + 500GB Seagate SATA2 NCQ 8MB & 32MB | 2GB OCZ PC3200 DC (2.5-3-2-5) | GeForce 6800GS 256MB | Sound Blaster Live! 5.1 | Razer Copperhead Laser Mouse | Standard M$ 101/102 Keyboard | Windows XP Pro SP2

BWX

I like the pretty colors.

__________________

Storm8873

I've tried Diskeeper 11 out and it seemed pretty good but I thought I'd give PerfectDisk 8.0 a go. It trashed the boot sector of my hard drive when it did the offline scan. I tried all the fixed on the net using the recovery console (fix boot, fixmbr, copy the ntldr files), but in the end I had to reformat and start again. After reading that a few other people had similar experiences, I thought that PD definitely has a problem with the offline scan.

Ciao ...

Robje

After reading this thread, which was very informative, I've dicided to go and test both products myself and comparing them to the standard features Windows XP offered. Using, what I perceive as, normal everyday computer usage and a set of fixed scenarios.

The computer used for these test was a AMD X2 4600+ with 2Gb of RAM and a Raid 0 array consisting of two 160Gb Sata hardisks.

Before I started the test, I made a ghost image of the PC. This image was used everytime a new product was installed to be tested, so as to create a identical starting point. The PC was used for 2 weeks on end, with my kids playing games, me doing work on various files using Autocad, Office, and some Videoediting software. During this period the partitions ( C- Drive (Windows XP X32): 40Gb (20% Free-space), D-Drive (Games): 120Gb (54% Free-space), E-Drive (Data): 150Gb (10% Free-space)) where defraged every monday, thursday and sunday, using the recommended setting for each product and if available automatic online defragmentation was turned on, again using recommended settings. After 2 weeks of PC use, a final defragmentation was done and various aspects where tested.

I approached this test as being a regular customer (in fact I,m a IT technician for over 10 years now working with various companies, and developers), meaning I used the information I could find on the various vendors websites for settings and configuration.
The timings where done using a proffesional stopwatch. I intentionally did not use any fancy timing tool, for a 'normal' non-techie customer wouldn't use such software.

The strong-point of both products, or the key-selling point for that matter, is a drastic improvement in the speed of your PC. As a regular customer does not care about the dynamics involved, but is just interested in the results, I tested the following:

Boot-time until logon-screen
Time after logon until desktop is presented without the hourglass cursor
Time to open a large moviefile for editing
Time to save a large moviefile
Response\boottime of various programs


After testing one product (and taking avarage times of 10 trials) the PC was cleared and the ghost-image was restored, before installing the next product to be tested.

And after 6 weeks of testing, what did I find out?

Well first, that my kids hate me for reseting there save-games every two weeks, and that I've fallen behind more then 6 weeks on social life, the following:

There is not much difference between the programs in respect to speeding up the PC. At best there was a mere 2 second gap between the products at boot-time or when saving a large-file. The claim of speeding up your PC is at best a nice fairytail (When compared to the standard tools available in Windows XP).

True, if your PC is cluttered with fragmented files, you can speed it up a lot by defragmenting your harddrives, but they don't do any wonders over the standard defrag tools incorporated in Windows XP. However, in all fairness, with both tools it is easy to automate the process and forget about it. Wheter you want it done real time, or sheduled, depends on your personal preferences.

As for myself, I was flabbergasted by the results, for I've been using and promoting 3th party defrag tools for as long as I can remember. Sure it makes defragmenting a bit easier than using the standard XP feature, but that's about it. For myself I can no longer justify the expense. Both products work, but their marketing claims are an understatement at best.

Is the expense for the products justified? That's up to the individual user to decide. For normal everyday PC use, coupled with normal PC maintenance, if speed is your main concern, I think not. If easy of use and\or automation is your goal, then maybe.

P.S. I congratulate the representatives of both vendors for taking an active part in this discussion! It is not something you see everyday.

For use in a server environment both vendors offer products that are vastly superior to the incorporated tools of your operating system. But that's a whole different ballgame all together!

BWX

Did you measure how long it took to defrag when doing it manually?

I've never measured but I would bet Perfect Disk is at least 2 times faster than built in defragger- that comes in handy when defragging a 160 and 500GB HD.

__________________

buddybird

Whew! For a while, I thought I was going crazy. PD 8 blew out my boot sector at least 3 times before I realized it wasn't me...

mike2h

ive been having probs with pd8 offline to. get weorh errors that rebbots fix but it is still a pain & troublesome.
this my 3rd upgrade of pd & have never had any issues til now(on 2 comps). hope this isnoy a trend & that they fix soon. otherwise this will be my last purchase/upgrade of their product.

__________________

BlueMak

This looks like the best place to ask, regarding Diskeeper 2007. What are the differences between the versions? Apart from the price, I can't find any list of features comparing the different versions. I am thinking of getting the Home but...

__________________

The people who are regarded as moral luminaries are those who forego ordinary pleasures themselves and find compensation in interfering with the pleasures of others(Bertrand Russell)"You go into Afghanistan, you got guys who slap women around for five years because they didn't wear a veil,You know, guys like that ain't got no manhood left anyway. So it's a hell of a lot of fun to shoot them." - Lt. Gen. James N. Mattis
This is slavery, not to speak one's thought. [Euripides-The Phoenician Women (c.411-409 B.C.)] http://www.macedonia.info/FALLACIESANDFACTS.htm Sic semper tyrannis.

Xelkos

Have seen this on my machine too, when I was comparing Perfect Disk to Diskeeper. Acronis TrueImage's Emergency Rebuilder loads on my hard drive, where I could press F12 and recover my OS from the last saved build.

Was pondering Perfect Disk for it's price, and that it claimed to have a free space optimization tool. The duplicate file finder I thought was a bit extreme for a defragmentation tool, that would be more utility software in my opinion.

Still, the price was better.


Tried Diskeeper 2010 and it was able to do the boot-time defragmentation, thus was the decisive factor.


Been using Diskeeper since version 10. Had Professional 2008 upgraded by their elevated solutions rep (Fields) when I discovered some hard drive issues. Feedback landed me a copy of Premier Pro 2008.
Yet was unable to upgrade to Pro Premier 2010 due to being out of work (currently); and the price was indeed a factor & why I was looking for a more affordable alternative
Luckily I found the older Professional build I had. Bought the $40 upgrade. Asked Diskeeper's Technical Support for another upgrade option - since Professional 2008 wouldn't install, was hard to "upgrade" since it would not install. Diskeeper got back to me the next day and sent links for the full version.

Diskeeper might not be the more affordable solution out there, but it works, and works quite well.
On my iPod 160 it defragments rather slowly, using an older 2.5"x2.5" heatsink that I set it on to keep the internal drive from overheating.
Diskeeper reported 45 fragments prevented, and 129 fragments eliminated on the iPod, @ 40.17 GB of music.

My 2¢ on iput.

PS: Photos of the iPod 160 & heatsink that I used.

BWX

LOL, it's funny you used that old pic, I haven't tried an offline defrag with PD since then and I probably never will again. Just way too dangerous.

__________________

Chaos

If you contact PD support there may well be a registry entry to solve that offline defrag problem. Im quite sure i went through that in the past

__________________

BWX

The last time I remember trying to fix that it was an incompatibility with deamon tools or similar type programs that use a form of a rootkit to function. AFAIK there was never a solution, but I haven't looked in a while.

__________________

Xelkos

Yeah, guess it is funny now that it's not 4:30 AM. Did not even look at the dates, and strange that the issue exists with the current release of PD.

As for the other fellow, I doubt Rexo would give support to a trial version. Still, Dk gets the job done without needing support.