The big XP defragmenter test

[BloodySword]

Oh I see.

But with SuperFetch disabled and keeping it optimized through MyDefrag the f***n interrupts and HDD cooking attempts will be out . The optimization is done when I WANT IT. Not when Windows thinks it would be the time to... This is what I hate on that. SuperFetch cooks my HDD everytime I work with VERY big images (scans of drawings of my girlfirend :'>) After GIMP releases some memory, SuperFetch cooks my drive. And the System is in this time VERY VERY slow. I hate it. I want it to be REMOVED in Windows 8. I also bought 2 500GB Hitachi TravelStar 7K500 500 GB for my Notebook and the other as externals drive because the 320 GB interhal HDD is a bit worn out... (Performance is degrading day by day, but no read or write errors, just SLOWER. Thanks SuperFetch... QQ)

So I think in this case Windows 7 tests would be very interesting. Vista can be skipped, its almoust the same I think, although SuperFetch is MUCH MORE AGRESSIVE - But this has nothing to do with the defragmenter of Windows 7 and Vista.

2014? That's werid. I read about 2011. Nevertheless is Windows XP really old (but somwhat faster as Windows 7 will ever ever be).

[Darlis]

But with SuperFetch disabled and keeping it optimized through MyDefrag the f***n interrupts and HDD cooking attempts will be out

You're taking about Vista don't you? I thought we already agreed that the Vista version is really bad.

I will run all the tests with SuperFetch enabled. Of course, like in the XP test, there will be a test without it, and after testing a few defragmenters, I think I'll also test the winner without SuperFetch.

2014? That's werid. I read about 2011. Nevertheless is Windows XP really old (but somwhat faster as Windows 7 will ever ever be).

2011 (or 2010) was the original expiry date but MS got a lot of complaints and had to extend the support.

[lh]

Very nice.  The only thing I would add in is maybe average time to complete a defragment pass.

[Darlis]

I've also recorded the time. You can find it in the respective thread of each defragmenter: http://hofmannc.de/forum/index.php?board=7.0

[BloodySword]

But with SuperFetch disabled and keeping it optimized through MyDefrag the f***n interrupts and HDD cooking attempts will be out ;)

You're taking about Vista don't you? I thought we already agreed that the Vista version is really bad.

No, I talked about Windows 7 SuperFetch, too. I tested it 2 months and I did not notice much improvement to Vista...

I will run all the tests with SuperFetch enabled. Of course, like in the XP test, there will be a test without it, and after testing a few defragmenters, I think I'll also test the winner without SuperFetch.

This sounds good. I gues it will be MyDefrag ;)

2014? :o That's werid. I read about 2011. Nevertheless is Windows XP really old (but somwhat faster as Windows 7 will ever ever be).

2011 (or 2010) was the original expiry date but MS got a lot of complaints and had to extend the support.

I don't wanna know how much Windows 2000 copies are still running out there. :o

[Bart Welson]

Nice test, but wondering how other free defraggers, thinking about auslogics or defraggler would compare in that list, especially because the latter has a boot-defrag option.

[Darlis]

I've already tested a program with boot time optimization: Perfect Disk
But the performance did not increase noticeably. Not even measurably with my tests. I guess with only 4 fragments the MFT was not fragmented enough so that a defragment would show any difference.

Btw. there is a poll on my forum where you can vote for the next defragmenter(s) I should test.

[Darlis]

I'd like to know how Monthly optimization results differ from using just Daily or Weekly optimizations in MyDefrag.

Just finsished testing the Weekly script: http://hofmannc.de/forum/index.php?topic=29.0

[Kasuha]

I'd like to know how Monthly optimization results differ from using just Daily or Weekly optimizations in MyDefrag.

Just finsished testing the Weekly script: http://hofmannc.de/forum/index.php?topic=29.0

Wow, thank you very much!
Apparently the performance gap between "sorted" and "filled" disk layouts is not that big. And IMO the main difference between "sorting" and "filling" scripts is that while weekly took almost as long as monthly script on first run, it is definitely going to finish much faster on second run (assuming periodic usage of scripts).
.. which confirms my theory that monthly script is only good for hardcore performance fanatics

[Darlis]

Apparently the performance gap between "sorted" and "filled" disk layouts is not that big.

Did you see the benchmarks, too? The Weekly was almost 4 seconds slower in the TV-Browser test. That is a noticeable difference.
And yes, the Monthly script delivers an (almost) perfect result. This is what some people want and so Jeroen has provided them this script.

It's also interesting to see that my Optimized Monthly has almost the same runtime and data movement as the Weekly but delivers even better results. 

[Kasuha]

Did you see the benchmarks, too?

Of course I did.

The Weekly was almost 4 seconds slower in the TV-Browser test. That is a noticeable difference.

This is the greatest difference in all of the results. And if I assume I start such a program two times a day every day over a month the total difference is 4 minutes. Not enough excuse for running the Monthly script for another 30 minutes in my eyes even if I only count the time spent on the optimization, not wear of the drive.

It's also interesting to see that my Optimized Monthly has almost the same runtime and data movement as the Weekly but delivers even better results. 

The difference is in second run. Even after a month I am sure the weekly script would be still much faster than monthly.

[Darlis]

And if I assume I start such a program two times a day every day over a month the total difference is 4 minutes. Not enough excuse for running the Monthly script for another 30 minutes in my eyes even if I only count the time spent on the optimization, not wear of the drive.

If you're eagerly waiting for a program to start, 4 seconds can be long; time is relative.
I've never written that you have to use the monthly every day. This is just the first run. Of course, the Daily and Weekly will finish much faster than the Monthly in the second run.

[BloodySword]

@Kasuha: You can not generalize it. The first programm is taking 4 seconds longer, the other program does even not respond for MINUTES (that happend to me with daily script and the program is GIMP x64 + Python [and the reason was NOT the building of fontcache!]).

I upgraded to a 7200 RPM 500GB drive for my notebook. This drive is increatable fast: ~130 MB/s sequentially read and write, ~50 MB/s random read and write with one 4K cluster.

I personally run the monthly script every two weeks, sometimes more often, sometimes more seldom. It depends on subjective performans degration. I will never use daily and weekly again for system drive. It is really dissappointing. Sometimes I use defragment only between monthly runs.

Data disk is optimized using Weelkly approach (renamed to Monhtly) every few months. Sometimes I use defragment only to optimize the performance a bit.

[Kasuha]

If you're eagerly waiting for a program to start, 4 seconds can be long; time is relative.
I've never written that you have to use the monthly every day. This is just the first run. Of course, the Daily and Weekly will finish much faster than the Monthly in the second run.

8.6 seconds sounds like a hell lot to me if I had to be eagerly waiting for that so if I'm patient enough to wait for so long why should I get pumped up in additional 3.3 seconds?

@Kasuha: You can not generalize it. The first programm is taking 4 seconds longer, the other program does even not respond for MINUTES (that happend to me with daily script and the program is GIMP x64 + Python [and the reason was NOT the building of fontcache!]).

I have no idea what do you do with your Gimp (it's never taking more than a few seconds to start on my machine) but I am sure just simple putting its files together in a separate zone would solve that single problem almost as well as sorting the drive, but at a cost of much less disk operations on optimization.

Sorting the disk is no doubt easy and provides good result but sorry, IMO it is overkill, takes long and wears out the disk, too.

[BloodySword]

I have used daily once and after that GIMP tooked a long time but it was not hanging at "Generating font cache", it was hanging while loading the plugins (only standard Plugins are installed). And not only that: Boot time increased (not much) and some other programs ware a bit slower, too.

When I use monthly, GIMP starts in a few seconds, even if I don't run monthly for some weeks.

Monthly will never run longer than 45 minutes on my Laptop with the 7200 HDD and there is 27 GB occupied space Windows 7 and programs and pagefile.sys (3GB) (Partition is ~56 GB)

Pagefile is placed on the very beginning and wrapped around some unmovables, then the normal script layout is following.