Experiment with Diskeeper's IntelliWrite (from PerfectDisk Blog)

Experiment with Diskeeper's IntelliWrite (from PerfectDisk Blog)

Fragmentation prevention with IntelliWrite -- say hello to reality

http://blog.perfectdisk.com/blog/all-things-perfectdisk--and-defrag/0/0/fragmentation-prevention-with-intelliwrite-say-hello-to-reality

 

Say hello to a nice advertisement I don't need graphs and partial screenshots from competitors to tell me Diskeeper is working perfectly. The real-time defrag is keeping my system purring and, as far as resource usage, it barely makes itself known in my experience.

 

Commercial Win Defrags: Talk BIG, Do small.

Freeware MyDefrag FTW!

By the way, PerfectDisk is the best of the commercial ones. At least it doesn't make useless features and doesn't create resource hogs.

Diskeeper makes useless features, as shown by this PerfectDisk dev's study, and then criticize competitors (YES, DISKEEPER DID IT FIRST):

http://www.diskeeper.com/blog/post/...-prevent-fragmentation-before-it-happens.aspx
X
http://blog.perfectdisk.com/blog/al...ention-with-intelliwrite-say-hello-to-reality

WAR! lol, serious: A discussion is going on between the 2 companies.

 

Like many tests they are not very helpful without any measurements taken over time. That is after 1 day, 1 week, 1 month of real world use how do different defragmentation tools compare in terms of fragment elimination, refragmentation and efficency of defrag and more importantly measurements of performance losses.

Relative to this test, maybe Intelliwrite appears to have bad free space fragmentation, but if it does not affect performance and does not get any worse with use and time compared to NTFS (which we know does gradually get worse and worse) then its actually superior to the vanilla NTFS drivers.

All I will directly comment on intelliwrite is that there are many Linux filesystems with superior fragmentation and performance loss resistance to NTFS, so there is room for improvement in the NTFS allocation algorithms. Some of the linux file systems are better by grouping together related files (could be by folder, age) and leaving some but large chunks of freespace so that changes to existing files and new related files can be located nearby.

So all I will say there is potential for intelliwrite to improve on NTFS's file placement algorithms.

 

Well, i did my own test. I used Diskeeper 2 or 3 weeks, now i installed PerfectDisk 10 and did an analyze on my HDD. I got the same results.

I think they are not lieing, they just think in another way of defragmenting.

2 things i dont like in Diskeeper:

1. very-very expensive
2. they think SSD needs defrag(they call it HyperFast)

I read a few complains about reduced HDD writing speed when they use IntelliWrite.

 

I tried the new version of Diskeeper shortly after it came out. It got removed quickly after I discovered that it did not play nice with FD-ISR... lost all my snapshots. FD-ISR users beware!! See here.

Ironically, my laptop seemed faster for the short time Diskeeper was installed but I am not giving up FD-ISR to continue using this,

 

1 thing I don't like in defraggers - they're snake oil. I remember doing defrags with Norton Utilities on my dad's 286 20 years ago, pretending that it made a difference. I really can't do that anymore. I haven't used a defragger in years - what I learned from that experience is that I should never again use one.
What I find amusing is new technical terms they coining out every once in a while and the tech blabberings ("benchmarks") they come up with to support the new terms and give them some meaning. Really a pathetic way to squeeze some cash from end-users.

The last I looked (which, I admit, was over a year ago) it had a bunch of services (not 1 but 2 or 3) running. It was off the system in a blink of an eye and I haven't looked at it since. That's how much I liked it.

The only way to do a proper defrag is to uininstall FD-ISR, defrag and then install FD-ISR again. You can defrag with FD-ISR installed (with some defraggers), but all you're doing is just fooling yourself.

 

Have you tried to exclude $ISR folder (DK configuration Properties --> File Exclusions) ?

 

I did not because I remember that was supposed to have become unnecessary with one of the releases of FD-ISR.

 

I would agree with you to a point. I have noticed an improvement in boot time by defragging system files with either PerfectDisk or Puran. Other than that I never notice performance gains. Diskeeper's marketing sucked me in to give it a try.

 

"Useless" is opinion only, so is "best of the commercial ones". You saying "Diskeeper did it first" sounds like a small child trying to excuse his or her behavior because "they started it!". Criticism of either company is not the subject of the thread, but whether IntelliWrite does its job. On my system, it does just that. The program itself uses 1 process "DKService.exe", coming in at a whopping 7Mb of RAM usage, period.

As with every program in existence, user experience may vary. But, I can't agree with the "snake oil" statement. For as long as I've been using computers, a few weeks worth of not defragging has proven time and time again to slow down my system.

 

one major flaw of diskeeper is that it won't tell you at a glance what the current defrag status is. Even though the program is always doing a defrag on the fly. It gives you alot of useless information, but leaves out what you really want to know. And on large drives it doesn't show you a realtime defrag in action which even the free mydefrag does (it claims it slows down the defrag) some of us like to see the little blocks of data move around during the defrag.

Perfectdisk does give you a current defrag status on the main screen. That way you know if it's working or not. The only problem with perfectdisk 2009 (the demo I tried) it wouldn't installed on my xp gaming computer which I use daily (it installed on another computer).

I like diskeeper always defragging on the fly, but now reading of the lack of performance gains I might find something else.

 

ive always wondered which out of these 2 big-players, was the better.....

i guess, we still dont know, seems to be more about opinions rather than statistical facts.

 

Not in the case of this IntelliWrite feature. And the best of the commercial ones is only an opinion, but it was based on real facts.

Oh well, sometimes the small child is right and those who really started are punished.

More like: criticism of either company's product features. BTW, IntelliWrite is useless not because it consumes lots of RAM, it's more about the BIG I/O activity Diskeeper generates over time, Diskeeper's IntelliWrite feature adds a 3rd party file system filter to the NTFS file system that "creates a situation where resources are consumed in order to prevent file fragmentation - but which results in the need to consume yet more resources to clean up the free space fragmentation issues it leaves behind."

 

Why is RAM usage so important in 2010? Almost every computer has 1GB RAM or more. I think the CPU usage and the file read/write speed(overall perfomance) are more important.

 

ram usage isnt,

the thing I want to know is: which defrags and places the files better for performance?

i know diskeeper has used 1min 11 secs of cpu compared to PD's 11 seconds, a full minute more.

Im not sure which to go for....

 

I get that, and yeah, I oversimplified. RAM is still important to me though, I certainly don't think a program should use more than is absolutely necessary. All that being said, I just don't see this "huge CPU usage" that's being complained about, at least not here. I'm no defrag expert, so whatever is up with this IntelliWrite feature, *shrug*, but, I go by what I see, and, so far at least, what I see is performance gain and the need for manual defragging gone.

Edit: Just for kicks, I ran a trial of PerfectDisk on a similiar setup to my main system. PD lists 3 running processes compared to Diskeepers single process. I'll be running PD one on system and Diskeeper on the other for a few days to see which one really does work better.

 

The method used by PerfectDisk (sorting files by modification date) is good to minimize re-fragmentation but a bad one to gain performance - the oldest files will be placed on the fastest disk area.
The way it is done by O&O Defrag 12 (Zones and resorting files by name & access) seems to be the better approach.

Are there other opinions?

MyDefrag is no alternative for me because of missing boot time defrag (a must have for me).

 

If I install an AV, firewall, couple of browsers, plugins and players and go wasting my time on various internet boards, then no, resources aren't of much importance to me.

Any serious productivity, and I certainly don't want a disk-munching process doing "IntelliWhatever" in the background. I need my HD, RAM and CPU as quiet as possible.
So it's not just about the RAM but the whole concept of resident defraggers.

Alright, I can understand on-demand defrag every once in a while (say, monthly) if it gives a warm fuzzy feeling and/or 2% improvement. But to me, defragger is more of a "dancing color boxes" gadget than anything else.

 

Well...i prefer PerfectDisk, im using since version 5 or 6.. i don't remember.

Download the trial from Diskeeper and PerfectDisk then decide. But if u choose PerfectDisk, wait until PerfectDisk 11 released. PerfectDisk 11 beta testing started today.

 

Which is better for overall performance - reduced fragmentation VS placing files on the fastest part of the disk (assuming it actually is) ?

 

As a vendor (Diskeeper Corporation), I certainly appreciate everyone taking your time to discuss these products.

I'd like to make a few clarifications, and am dumping multiple answers (from questions/comments in this thread) into one post. Sorry for the novel.

The purpose of the Diskeeper Blog post, of which I am the author, was not to criticize anyone or, LOL, incite some "blog war" . Note that I did not name anyone (nor do we ever). The purpose was to distinguish, for users, between file placement algorithms designed to slow file re-fragmentation and proactive technology that prevents file fragmentation. They are different.

It is unfortunate that the article has been misinterpreted by readers and competitors. I'll be more careful with my wording in future.

Resource usage:
As for resource usage, InvisiTasking and IntelliWrite DO NOT impact the system. IntelliWrite's use of resources, in doing a better job of preventing fragmentation than the native NTFS functionality, is negligible. The benefits of IntelliWrite far outweigh the requirements.

The fact that I/O Other bytes accumulate for the DKservice.exe process is irrelevant; it does not impact the system. Yes, I/O Other bytes can represent impact (e.g. RDP), but for Diskeeper I/O Other bytes largely represents InvisiTasking, using idle resources, determining whether Diskeeper is allowed to do any actual DISK reads or writes. Poorly timed defrag-related Disk I/O can and will impact the computer - what InvisiTasking is designed to avoid. I/O Other bytes is simply a statistic that may be noteworthy, nothing more. To infer that it's activity empirically impacts system performance would be misleading.

If you run performance tests and/or benchmarks (e.g. PCMark), you'll see that scores are higher when InvisiTasking and IntelliWrite are active. Whatever "resources" are used, are more than returned performance-wise (a net gain). I do not see benchmarks as irrelevant. They are rubber-meets-the road info that gets past theories, inferences and opinions.

SSDs:
Most SSDs do benefit from free space consolidation. Raxco has a similar statement: http://www.perfectdisk.com/support/kb/791. You can also visit OCZforums where users and, kudos to them, OCZ staff actually did tests: http://www.ocztechnologyforum.com/forum/showthread.php?t=64801&page=7

File Placement:
The I-FAAST feature (first released in 2005 - Diskeeper 10.0) places key files in the faster regions of an HDD(s).

Free Space:
Free Space fragmentation is not, in and of itself an issue. It's only an issue to the degree that it precipitates NTFS file write fragmentation. IntelliWrite addresses that at the "source". Therefore free space fragmentation is relatively moot, unless it gets severe. Also note that Diskeeper will consolidate free spaces, to the degree it is valuable.

Testing over a longer period:
An isolated case on an inactive and unused volume is not much of a real-world test. One quick real-world test anyone can run (i.e. with Diskeeper 2010 30-day trialware) is to clean up most of the pre-existing fragmentation on your PC. Then turn off IntelliWrite for a week. Then for the second week turn it on. Use any defrag tool to measure the number of file fragments that accumulate in each week. Or, if you have the defragmenter turned on (to do background/scheduled defrags), note how many less file fragments it has to remove in the week that IntelliWrite is active. And, remember the amount of free space fragments is irrelevant unless it translates to file fragmentation; and therefore performance loss. The only caveat is that the usage of the computer in the experiment should be fairly equal for both of those weeks.

Customers/Marketing:
The overwhelming majority of Diskeeper Corp business is working with IT Professionals at other companies/organizations, not selling to users. I'd imagine that is probably true for Raxco and O&O as well.

Please understand I'm not saying that to dismiss the home user (I'm a home user who buys software too), we value and care for every customer large or small. I make that point because we focus our ads/communication to IT Professionals. The "big talk" you may read is based on successes that many customers do get, not a guarantee for everyone. IT professionals are insanely busy people and the job of marketing is to encourage busy professionals to take the time to try the software. We believe once they try, they'll like it and want to buy it.

If the software doesn't live up to expectations/demands, you have the option to not buy it. Most every defrag vendor offers a free trial so you can make that decision yourself, using any standard/process you are comfortable with. Diskeeper has a 30-day unconditional money-back guarantee as well, should you choose to buy it before trialing. I'd imagine other vendors may have similar policies.

Cost:
Yes, Diskeeper has been generally been priced higher over the years, and our prices have stayed at a consistent level. Others have undergone price drops - likely in order to remain competitive. That says a lot about Diskeeper's continuing technical value in and of itself. Whether that extra hard earned money is worth it to you, is entirely your decision - based on what you value.

Hope this helps,
Michael Materie
Director of Products and Technologies
Diskeeper Corporation

 

One other technical bit I left out of the "resource usage" section:

For anyone curious, much of Diskeeper's I/O is done through IOCTLs (http://en.wikipedia.org/wiki/Ioctl), which gets logged as I/O Other bytes.