New optimization method

[Digiface]

So those certain files are moved, if i just run i.e. default action 3?

Another question: I last time ran that boot-time defrag, without any parameters and all the files were moved to the end of the disk... If i next time run just action 3, is those files going to be moved back?

[jeroen]

So those certain files are moved, if i just run i.e. default action 3?

Yes.

Another question: I last time ran that boot-time defrag, without any parameters and all the files were moved to the end of the disk...

That is not possible. If you run JkDefrag without parameters then everything is moved to the begin of the disk, not the end.

If i next time run just action 3, is those files going to be moved back?

Ofcourse.

[Digiface]

That is not possible. If you run JkDefrag without parameters then everything is moved to the begin of the disk, not the end.

But you said this in earlier message:

Run JkDefrag without parameters whatsoever to move them to the fast end.

[jeroen]

The begin of the harddisk is the fast end.

- If you run JkDefrag without any parameters whatsoever then it will move all files to the begin of the harddisk, which is the fast end.
- If you run JkDefrag with just the "-a 6" parameter and nothing else, then everything on the disk will be moved to the slow end, which is the end of the harddisk. The "-a 6" option is designed to move specific items to the end, for example "jkdefrag -a 6 c:\recycler\*". If you forget to specify items then everything will be moved.

[Digiface]

The begin of the harddisk is the fast end.

Ok. I misunderstood the meaning of end.

So those certain files are moved, if i just run i.e. default action 3?

Yes.

How about with action 2?

[jeroen]

How about with action 2?

Yes, action 2 in v3.9 also uses the internal list of SpaceHogs.

[Digiface]

Ok. Still got one question: If i want to keep files in the fast end, i always have run without parameters? Does 3.9 still perform actions 1-6, without any command parameters specified?

[jeroen]

Ok. Still got one question: If i want to keep files in the fast end, i always have run without parameters?

No. All "-a" options will optimize to the fast end of the harddisk, except option "-a 6".

Does 3.9 still perform actions 1-6, without any command parameters specified?

Yes. Steps 2...5 are now combined into 1 phase, so the program only counts to 3, but the same things happen.

[rdsu]

Jeroen,

JkDefrag is getting better every new version, and that is great for a free program...

Will be possible to have other method based on the last accessed time, to put the used files on the outer tracks (layout.ini and most used files) and the % of the files not accessed on the inner tracks. Something like the we can do with UltimateDefrag, where we can use the Pareto principle and much more...

Regards

[jeroen]

Will be possible to have other method based on the last accessed time

See the "-a 9" option.

[JDPower]

Will be possible to have other method based on the last accessed time

See the "-a 9" option.

Is it worth adding a note to the 'Sorted optimisation' section on the homepage, regarding the '-a 9' option regarding the fact that most tweaking programs disable the 'Last access time stamp' (This tweak here), which would stop this option giving any real optimisation? (or is that not what JKD uses to tell the last access time?)

[rdsu]

Will be possible to have other method based on the last accessed time

See the "-a 9" option.

I already saw that, but it doesn't make the rest...

[jeroen]

Yet another reason to stay away from tweaks...  The article you quote says it only applies to directories, and Microsoft Technet concurs. So the scope of this problem is very limited, only directories and only the "-a 10" option. JkDefrag will still place all directories together at the begin of the harddisk, but always in the same order. So I don't think it's a big problem, not important enough to bloat the docs with. Thanks for the input, though.

[JDPower]

Yet another reason to stay away from tweaks...  The article you quote says it only applies to directories, and Microsoft Technet concurs. So the scope of this problem is very limited, only directories and only the "-a 10" option. JkDefrag will still place all directories together at the begin of the harddisk, but always in the same order. So I don't think it's a big problem, not important enough to bloat the docs with. Thanks for the input, though.

Ok, thanks for the further explanation. Just thought it was worth mentioning just in case it was relevant.

[rdsu]

Jeroen,

Do you read the help file, or website description, of UD about that, and about their OptiSeek method?

Where you put the "layout.ini" files? In the beginning of zone 2?

And why not put the directories before or after MFT to improve the seek time?

From UD website:
Consolidate and Place Directories In The Optimum Position For Superior Performance
Another significant cause of reduced and sluggish performance is the scattering of directories/folders all over your hard drive in no particular order. It's as big a cause of reduced performance as fragmentation. When you access a file - hard drive needs to query the MFT (Master File Table), then the directory, then access the file. The MFT/directory transactions can often be multiple transactions before it finds the file it needs. The more scattered, the more travel, the slower the performance. UltimateDefrag takes all of your directories and puts them together and in the best position possible for accelerated performance.

Optimum Directory Placement And Consolidation
On NTFS formatted drives, a lot of dialog and data transactions occur between the MFT (Master File Table) and directories. Accessing a file often comes with multiple read/write head movement between the MFT, root directory and the subdirectory that the file is in. If your directories are all over the drive - a lot of overhead is wasted in accessing files.