JkDefrag 3.9 & Full Optimization (-a 4)

[dsage]

I noticed the documentation for version 3.9 omits the -a 4 option for full optimization.  Is this function still supported?

[jeroen]

The function is still in the program, but I've never been happy with it and I think I will take it out in a next version. The Windows defragmentation api on Win2K had big problems with it and would make a lot of fragmented files. The sorting optimizations are much better.

[dsage]

Which sorting optimization would you recommend in place of the full optimization?

[jeroen]

Which sorting optimization would you recommend in place of the full optimization?

I'm sorry but I have no recommendation. All five sorting methods have there advantages, see the Frequently Asked Questions, the "Running" chapter, the "Sorted Optimization" paragraph.

[dsage]

I understand your reluctance to provide a recommendation given the nature of the sorting strategies, and certainly the multiple options you offer are attractive to the power user, but this can also make it confusing for the "basic" user.  Just a thought.

My understanding is that most (all?) other disk optimization utilities place files on the disk according to usage.  Would the "sort by last access" strategy most closely mimic what other programs, such as PerfectDisk, are doing?

Also, I'm curious about the "sort by name" strategy.  Does this include the full path so that files in a given directory are grouped together, or does it just use the file name alone?

Lastly, the FAQ suggests a daily run with the "-a 3" option and a weekly run with the "-a 10" or "-a 11" optimization.  Do you see a significant difference between 10 and 11?  I'm trying to decide between the two.

Thanks for your help.

[jeroen]

I understand your reluctance to provide a recommendation given the nature of the sorting strategies, and certainly the multiple options you offer are attractive to the power user, but this can also make it confusing for the "basic" user.  Just a thought.

Optimization depends on usage, for example a server has totally different requirements than a desktop, or a notebook. And directories should be optimized differently than pictures, or databases, or archives. I really cannot give a general recommendation, it depends on how you use the computer.

My understanding is that most (all?) other disk optimization utilities place files on the disk according to usage.  Would the "sort by last access" strategy most closely mimic what other programs, such as PerfectDisk, are doing?

I have looked at the article you quote, but it is not clear to me what they mean. It could be "sort by last acces", yes.

Also, I'm curious about the "sort by name" strategy.  Does this include the full path so that files in a given directory are grouped together, or does it just use the file name alone?

Full path.

[dsage]

I'm running a standard desktop.  The FAQ suggests a daily run with the "-a 3" option and a weekly run with the "-a 10" or "-a 11" optimizations to minimize daily and weekly runtimes.  Is this the way to go?

I tried running with the "-a 10" (last change) option, figuring that moving the oldest files to the front would make subsequent runs faster (since old files that aren't modified shouldn't get fragmented).  However, I ran it several times in a row and subsequent runs were still slow (unlike with the "-a 3" or "-a 4" options where it will speed up if you run it several times in a row).  In fact, each time it started by moving blocks at the bottom of the screen, which makes no sense since the oldest files shouldn't need to be continually moved around.  Can you explain this behavior?

[jeroen]

This is just the first version of the sorting optimizations. The program does not check if files are in place already and will always rewrite all the files on the entire disk. I hope to add some code in the next version that will check if files are already in place and then skip them.

[Ratte]

The function is still in the program, but I've never been happy with it and I think I will take it out in a next version.

I am to leave the function nevertheless in the program.

The Windows defragmentation api on Win2K had big problems with it and would make a lot of fragmented files.

I see the problem regularly, but it concerns rarely more than 2 to 3 from  750 000 files in 15 000 folder. And these files have then also usually only 2 to 3 fragments. That is quite acceptable for me. There it disturbs substantially more that frequently different files cannot be moved, although no more program uses or since the start of PC was not at all used them.

The sorting optimizations are much better.

Those may be better, but they need to have also longer and for it more files to move.

[dust]

This is just the first version of the sorting optimizations. The program does not check if files are in place already and will always rewrite all the files on the entire disk. I hope to add some code in the next version that will check if files are already in place and then skip them.

can it be at the options a10 and 11 that the files are written to gaps which may fit and not to the end of the disk? that may explain why it hangs here at about 20% while step 3. sounds similar to http://www.kessels.com/forum/viewtopic.php?t=294

don´t know if they use the same options, so i posted here.

[jeroen]

can it be at the options a10 and 11 that the files are written to gaps which may fit and not to the end of the disk? that may explain why it hangs here at about 20% while step 3.

No.

[dust]

can it be at the options a10 and 11 that the files are written to gaps which may fit and not to the end of the disk? that may explain why it hangs here at about 20% while step 3.

No.

when it frees space at the beginning it does not write the files to the end to get free space, i have the impression it tries to write them to a bigger gap. 100% cpu usage without files moving and small not filled gaps at the end suggest that.

[Rash]

Although this is a little offtopic already, I have one question about optimizations too. I am thinking about doing different optimizations on the same partition and I would like to know how it would be like. Sorting by name on Program Files and Windows (hoping it would speed up loading times) and a simple optimization (a3) or by last change (a10) on Users (Vista's Docs & Settings) and Windows\Temp dirs. Wouldn't one optimization interfere with the other? That is, wouldn't an optimization undo everything the other one did? And if I run a a3 daily on my entire disk, would it undo the optimizations made by a7 and a10 on the other directories?

Ow, by the way, this program is awesome! Thank you very much.

[jeroen]

I am thinking about doing different optimizations on the same partition

The sorting optimizations will undo everything another optimization has done. The "-a 3" option however will leave most files alone and only defragment the fragmented files, and fill gaps with files from above. So, what you could do is use one of the sorting optimizations first, and then use "-a 3" every day for some weeks.

[Guspaz]

But, hard disks are a lot faster closer to the start of the disk. Without the '-a 4' option anymore, how do we get files to the faster part of the disk? Isn't this one of the main two benefits of defragmenting (apart from actually defragmenting the files)?

Would an equivalent now be to use '-a 5' (to force all files to the front of the disk) followed by '-a 3' to defragment any files fragmented during the moves?