No Space

[ti1358]

@Scannan:

The cluster size on my system partition is 4K, the partition is 32GB.

(32 * 1024 * 1024 * 1024) / (4 * 1024) = 8,388,608 clusters

 

SDefrag has (on my screen, by default) 102 x 12 = 1124 blobs

So each blob must represent 8,388,608 / 1124 = 7463 clusters

 

If ONE of those 7463 clusters contains a fragment of a file, then that blob is painted Red and it is not obvious that the other 7462 clusters are empty

 

There is no reasonable way around this problem ...If you had a drive which was exactly 1TB (after formatting) with 32K Clusters and each cluster were represented by a single pixel you would need ((1*1024*1024*1024*1024) / (32 * 1024)) / (1080 * 1920) = 17 1080p monitors to display the drive-map accurately !

 

Please let us not discuss drive manufacturers using 10^3 not 2^10 for marketting purposes.

 

I hope that makes sense.

 

@Melvin:

Yes, this conversation has developed in ways I too did not expect ...As you say, I solved my personal problem by using other software ...But it seems only polite to offer help understanding the problem (Eg. The maths above) in return for the warm reception and kind help I have been welcomed with on this forum.

 

The "Algorithm" (a function with decisions) may *choose* not to move files, but this is very different from "unmovable/Black disk clusters" ...As you say, everything on a HDD is theoretically movable, but restrictions apply ..EG. When windows is running you simply *can't* move the pagefile - hence the boot-time options in SDefrag - and hence them being Black when Windows is running ...EG2. If you move your Boot Manager (EG. LILO) overlay code your machine will not longer boot unless you tell the program in the boot sector to where you moved it's overlay code ...As there are hundreds of Boot Managers, this is an almost impossible task hence making this code "unmovable"/Black

[ti1358]

Re moving locked DLL's

 

This is a nightmarish subject which includes the concept of Shadowing, Mutexes and other more specific programatic considerations ...But yes, in general it is easier and safer to move system-locked files during boot - so you're probably right there - I had not considered that :)

 

One way or the other the proof will be presented when we can see the list of files marked as Unmovable/Black.

 

Ti

[Scannan]

Ti

 

Yes, it will be very interesting to see what the next version of SD will be like. There has certainly been a lot of food for thought presented in this thread.

[solbjerg]

Hi ti

By me if I count the squares in the program window I will in the default non-maximized window arrive at appr. 1000 squares, - if I maximize the window I will arive at appr. 3000 squares - my screen is a 17" screen.

 

I too sometimes (once in a long while) utilize the method of copying and then deleting and then pasting the copy in order to defragment some data

 

The interest in the methods applied in the inner workings of defragmenting is understandable for technical people - the parameters applied may be one of the reasons that SmartDefrag does the defragmentation quickly and this in turn may be why the developers does not want to divulge to much information about it. Just a guess mind you. :-)

The decision to keep the SmartDefrag a free program is in my view a very nice way to promote the company, and I understand why they do not want to give out too much technical information on the inner workings of the program.

Cheers

solbjerg

p.s. Left an Edit: in post #22

 

 

@Scannan:

The cluster size on my system partition is 4K, the partition is 32GB.

(32 * 1024 * 1024 * 1024) / (4 * 1024) = 8,388,608 clusters

 

SDefrag has (on my screen, by default) 102 x 12 = 1124 blobs

So each blob must represent 8,388,608 / 1124 = 7463 clusters

 

If ONE of those 7463 clusters contains a fragment of a file, then that blob is painted Red and it is not obvious that the other 7462 clusters are empty

 

There is no reasonable way around this problem ...If you had a drive which was exactly 1TB (after formatting) with 32K Clusters and each cluster were represented by a single pixel you would need ((1*1024*1024*1024*1024) / (32 * 1024)) / (1080 * 1920) = 17 1080p monitors to display the drive-map accurately !

 

Please let us not discuss drive manufacturers using 10^3 not 2^10 for marketting purposes.

 

I hope that makes sense.

 

@Melvin:

Yes, this conversation has developed in ways I too did not expect ...As you say, I solved my personal problem by using other software ...But it seems only polite to offer help understanding the problem (Eg. The maths above) in return for the warm reception and kind help I have been welcomed with on this forum.

 

The "Algorithm" (a function with decisions) may *choose* not to move files, but this is very different from "unmovable/Black disk clusters" ...As you say, everything on a HDD is theoretically movable, but restrictions apply ..EG. When windows is running you simply *can't* move the pagefile - hence the boot-time options in SDefrag - and hence them being Black when Windows is running ...EG2. If you move your Boot Manager (EG. LILO) overlay code your machine will not longer boot unless you tell the program in the boot sector to where you moved it's overlay code ...As there are hundreds of Boot Managers, this is an almost impossible task hence making this code "unmovable"/Black

[ti1358]

The formula to calculate the number of Clusters represented by a SDefrag "Blob" is:

ClusterCount = DriveCapacity / ClusterSize 
BlobSize = ClusterCount / BlobCount
-or-
BlobSize = (DriveCapacity / ClusterSize) / BlobCount

 

The drive Capacity is listed at:

RightClick Drive -> Properties -> Capacity

 

The cluster size can be found by:

WindowsExplorer:
 File -> New -> Text Document
DoubleClick "New Text Document.txt" to open it in Notepad
 Type "x" in to the file
 File -> Save
 File -> Exit
Back to WindowsExplorer:
 RightClick "New Text Document.txt" -> Properties -> Size On Disk

 

Because no file can be smaller than a single cluster, our new 1-byte file will use an entire cluster on the drive ...and therefore announce the Cluster Size

 

==

 

I have also just been reading about the USN in the "UltraDefrag Handbook" ...this is part of the NTFS journaling system ...If you have a number of unmovable chunks of disk you may find that deleting a recreating the USN will help consolidate those chunks:

 

Eg. Here is how I did it:

C:>fsutil usn queryjournal C:
Usn Journal ID   : 0x01cac3ccd11c06a0
First Usn        : 0x000000022f800000
Next Usn         : 0x0000000231a43f58
Lowest Valid Usn : 0x0000000000000000
Max Usn          : 0x00000fffffff0000
[color="Red"]Maximum Size     : 0x00000000020c0000[/color]
[color="Green"]Allocation Delta : 0x0000000000400000[/color]

C:\>fsutil usn deletejournal /n C:

C:\>fsutil usn createjournal [color="red"]m=34340864[/color] [color="Green"]a=4194304[/color] c:

I used a calculator to convert the hex numbers to decimal
I don't know if this was required, but it works.

There is no reason why SDefrag should not allow you to opt-in and perform this operation automatically.

 

UltraDefrag also openly publish their algorithms in the Introduction of the Handbook if you are interested to know more about the technical side of defragmentation ...although, imho, they are nothing special - the method is interesting, but the end result seems that it will be inferior to the SDefrag end result.

 

Ti

[Scannan]

ti

 

Great info. Much appreciated. Will play with it.

[Melvin_Deal]

Hi Ti.

 

Who is...

our

?

 

And we...

we

Instead of ti1358?

 

Sincerely,

-Mel

Live long and prosper!

[ti1358]

LOL :D

 

Who is... "our"? And we... "we" Instead of ti1358?

Does one/Mel mean: "Who is "our" and "we" (in place of ti1358)?"?

...ooohh, nested quotes and valid quadruple punctuation - spot the bored techie! ;) LOL

[solbjerg]

Hi ti

You left a good explanation on why the interface window in the program can show that it is totally filled with coloured squares while the report says that you still have some free space.

Thank you for that!

Cheers

solbjerg

 

 

LOL :D

 

 

Does one/Mel mean: "Who is "our" and "we" (in place of ti1358)?"?

...ooohh, nested quotes and valid quadruple punctuation - spot the bored techie! ;) LOL

[solbjerg]

Fragmentation

 

Hi

Some general information:

http://en.wikipedia.org/wiki/Slack_space

 

Cheers

solbjerg