Intro Download and install Frequently Asked Questions Tips and tricks

Homepage







© J.C. Kessels 2009
MyDefrag Forum
September 01, 2014, 07:29:09 pm *
Welcome, Guest. Please login or register.

Login with username, password and session length
News:
 
   Home   Help Search Login Register  
Pages: [1]
  Print  
Author Topic: Best place for MFT zone?  (Read 7686 times)
Raziaar
Newbie
*
Posts: 2


View Profile
« on: January 08, 2008, 10:57:21 am »

I know we cannot move the MFT zone or anything like that.

I just reformatted my computer, and before reformatting my MFT was near the middle to end portion of my hard disk. To me that seemed the best place since it let all the space for my files be taken up at the beginning.

Now that I reformatted, I see that my MFT zone is at the beginning of the disk(bottom of screen in JKdefrag). Is this ideal or less than ideal?
Logged
jeroen
Administrator
JkDefrag Hero
*****
Posts: 7220



View Profile WWW
« Reply #1 on: January 08, 2008, 11:09:31 am »

The beginning of the disk is the fastest part of the disk, and the MFT is accessed very often, so that is a good spot. An even better spot would be the average position of the disk heads, which is not the middle of the disk, but the middle of the data (actually it depends on things like how the data is optimized and how the speed of the disk diminishes towards the end).
Logged
Raziaar
Newbie
*
Posts: 2


View Profile
« Reply #2 on: January 08, 2008, 11:24:24 am »

So... in other words I'm good with where it is? Thanks!
Logged
svhb
Newbie
*
Posts: 3


View Profile
« Reply #3 on: January 09, 2008, 10:58:07 pm »

I'm confused a little bit now.

Why is the beginning of the disk the fastest part? It is only because there is most of the data over there (or the most often accessed data = the operating system = at the begginning because it's installed first).

If the MFT table is accessed a lot, and not movable, should a possible strategy not be to move the other files 'around' the MFT? This should generate a disk like this : center - empty - data - mft - data - empty.

Or is this a completely wrong idea?


Stefaan
Logged
jeroen
Administrator
JkDefrag Hero
*****
Posts: 7220



View Profile WWW
« Reply #4 on: January 10, 2008, 04:24:51 am »

Why is the beginning of the disk the fastest part?
See the "see also" section on the JkDefrag homepage for a link to HDtune, a small free program to measure the speed of your disk. The beginning of the disk is faster for physical reasons, it has nothing to do with data.
Logged
Esprit
JkDefrag Hero
*****
Posts: 70


View Profile
« Reply #5 on: January 10, 2008, 07:05:22 am »

Why is the beginning of the disk the fastest part? It is only because there is most of the data over there (or the most often accessed data = the operating system = at the begginning because it's installed first).
Beginning of the disk is fastest because the heads are on the biggest circle. Disk is rotating in constant (angular) velocity so the linear velocity (under heads) is biggest thus the fastest reading.
Logged
svhb
Newbie
*
Posts: 3


View Profile
« Reply #6 on: January 10, 2008, 02:25:57 pm »

Why is the beginning of the disk the fastest part? It is only because there is most of the data over there (or the most often accessed data = the operating system = at the begginning because it's installed first).
Beginning of the disk is fastest because the heads are on the biggest circle. Disk is rotating in constant (angular) velocity so the linear velocity (under heads) is biggest thus the fastest reading.
This suggests that there is more information on the outside of the disk than on the inside, wich is not true I think
Logged
jimbo
JkDefrag Hero
*****
Posts: 84


View Profile
« Reply #7 on: January 10, 2008, 04:22:29 pm »

Actually, it is exactly true. For several years now, there have been a variable number of sectors on the tracks of a disk, with more sectors on the outer (beginning) tracks than on the inner (end) tracks.

This has been happening with SCSI disks for a lot longer than IDE, but even with IDE (PATA and SATA) disks it is pretty much normal now.

The cylinder/head/sector counts have been fiction for a long time since they were simply to enhance compatability with older machines from pre-LBA (linear block addressing) days, but internally to the drives, pretty much all access is by sector number (LBA) now. When you factor in that modern hard drives are also capable of remapping individual defective sectors to other areas of the disk, the LBA numbering over c/h/s becomes even more essential as the only real way of taking to a part of a disk.

So, yeah, that is a large part of the reason why the outside of a disk is faster than the inside.

If you're interested in a reference, try http://www.dewassoc.com/kbase/hard_drives/hard_disk_sector_structures.htm
« Last Edit: January 10, 2008, 04:26:53 pm by jimbo » Logged
svhb
Newbie
*
Posts: 3


View Profile
« Reply #8 on: January 10, 2008, 06:55:15 pm »

Good to know

Thanks
Logged
poutnik
JkDefrag Hero
*****
Posts: 1112


View Profile
« Reply #9 on: June 08, 2008, 01:29:20 pm »


If the MFT table is accessed a lot, and not movable, should a possible strategy not be to move the other files 'around' the MFT? This should generate a disk like this : center - empty - data - mft - data - empty.

Or is this a completely wrong idea?

This part of a question was not answered....

My note and ideas :

Fastest transfer disk area and fastest access disk area can differ,
if MFT is located somewhere in the middle of data for whateven reason.
E.g. some partition resizers place it there.

Beginning of the disk can be in these cases slow location for folders
and frequently accessed small files.
Access time is a bottleneck here, not transfer bandwidth. 
Relation of their differences is even worse.

To borrow arabic sayings:
"If Mohamed ( MFT ) cannot go toward the mountain ( folders (AND frequent small files )) ,
 the mountain has to go toward Mohamed.    Smiley

 
« Last Edit: June 08, 2008, 05:22:33 pm by poutnik » Logged

It can be fast, good or easy. You can pick just 2 of them....
Treating Spacehog zone by the same effort as Boot zone is like cleaning a garden by the same effort as a living room.
jeroen
Administrator
JkDefrag Hero
*****
Posts: 7220



View Profile WWW
« Reply #10 on: June 08, 2008, 05:27:10 pm »

Access time is a bottleneck here, not transfer bandwidth.
Yes, access time is more important than transfer bandwidth for small files. But I'm not sure if I agree with your conclusion that small files should therefore be as close to the MFT as possible.

Firstly, there's the Windows disk cache to consider. It's main focus is to minimize MFT access. Without the cache, the harddisk heads would have to move from the MFT to a file, and then back to the MFT again for the next file. The cache makes accessing the MFT unnecessary, the heads can move to the next file right away, for the vast majority of file accesses. I have no hard numbers but my feeling is that more than 90% of MFT accesses are prevented, depending on how the computer is used. So, the distance between the small files is far more important than the distance of the small files from the MFT.

Secondly, not only the transfer bandwidth decreases from beginning to end of the harddisk, but also the access time. On an average harddisk the access time at the end of the disk can easily be 300% slower than at the beginning. You can measure this for yourself with a tool such as HDTune, which I mentioned before in this thread. The technical reason for this is that tracks contain less data near the end of the disk, so the heads have to skip more tracks. Please note that I am not talking about access time from the beginning of the disk.

So, distance between files is more important than distance to the MFT, and the beginning of the disk has faster access time. I think that small files should be placed at the beginning of the disk, irrespective of the MFT. It would be even better if the MFT is also at the beginning of the disk....
Logged
Lundholm
JkDefrag Hero
*****
Posts: 208



View Profile
« Reply #11 on: June 09, 2008, 07:01:04 am »

If the MFT table is accessed a lot, and not movable, should a possible strategy not be to move the other files 'around' the MFT? This should generate a disk like this : center - empty - data - mft - data - empty.

Or is this a completely wrong idea?
Good idea, and good ideas tend to have this extra quality that somebody else already thought of it. Smiley

You may want to look at this: http://www.kessels.com/forum/index.php?topic=924.0
Logged

"O, there has been much throwing about of brains." -- Guildenstern{alt. Gyldenstern[alt. Gyldenstjerne(anc. Gyllenstierna{knight of Lundholm})], knight of Hamlet}.
Pages: [1]
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.5 | SMF © 2006-2008, Simple Machines LLC Valid XHTML 1.0! Valid CSS!