-
We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.
Poll: Windows2000- NTFS or FAT32?
- Thread starter zippy
- Start date
Mean MrMustard
Diamond Member
Definitely NTFS for Win2000.
<< Besides, security, what are the advantages to NTFS? >>
NTFS is a journaling file system. In a nutshell that means there will be less of a chance of data corruption. NTFS also has built-in compression, encryption, as well as other little goodies.
Make sure when you reformat the Master File Table is at the front of the disk or you are going to have MAJOR performance problems when the disk gets fragmented. I think the Win2k installer does it by default but I never could get a clear answer on this.
Despite what anyone says, FAT32 is faster than NTFS. The fact that FAT32 is a stupid little lightweight filesystem makes it faster. There are no file security descriptors to check on file access, thus the speed advantage over NTFS.
I personally will not use NTFS on my boot partition at home in case I need to install Win9x or boot to a DOS command prompt from a floppy.
danielshoes
Senior member
I placed this post at 24/01/01 and asked for benchmarks. Somebody answered me with this very good review below:
- - - - - - - - - - - - - - - - - - - - - - - - - - -
FAT-32 ON WINDOWS 2000: PENNY-WISE AND POUND-FOOLISH?
by John L. Joseph, Diskeeper Development Section, Executive Software
Editor's Note: If someone is described as "penny-wise and pound-foolish", they are thought to be too careful about very small details and not able to consider larger, more important matters. (Cambridge International Dictionary of English). The expression comes from English (UK) currency, the penny being the much smaller unit and the pound (Sterling) being the larger.
-------------------------------------------------------
As more and more individuals and companies convert to Windows 2000, it's becoming a more and more common question to hear: "Should I use FAT-32?"
It is even more important to answer when I'm told by one user that he just received a desktop machine with Windows 2000 preinstalled on a 20GB drive under the FAT-32 file system, and another user tells me he just received his new laptop with Windows 2000 preinstalled on a 30GB drive under the FAT-32 file system.
Because of other problems I'd seen with the way mail-order computer vendors configured pre-installed systems, I knew I'd have to check it out.
The first place I went to was of course the Microsoft Knowledge Base*.
In article Q100108, I found this discussion:
"Disadvantages of FAT"
Preferably, when using drives or partitions of over 200 MB the FAT file system should not be used. This is because as the size of the volume increases, performance with FAT will quickly decrease."
Wow! That's a bit of a revelation for most computer users I know.
I then found this in article Q154997:
"For most users, FAT32 will have a negligible performance impact. Some programs may see a slight performance gain from FAT32. In other programs, particularly those heavily dependent on large sequential read or write operations, FAT32 may result in a modest performance degradation."
The obvious thing to do was to quantify what "modest" really meant.
I took a 866MHz Pentium III Windows 2000 machine with a DMA-66 IDE hard drive, and carved a 2GB volume out of the hard drive. I then created a program which would place 5200 files of known, predicted size on that volume, filling up the volume till it was 24% free, using all three file systems. I recorded the number of wall-clock seconds it took to complete the job. Here's how it came out:
- Under FAT-16, under which I could only use a 64K-byte cluster size, the job completed in 1,682 seconds.
- Under FAT-32, with a 4K-byte cluster, the job completed in 2,472 seconds; with a 1K-byte cluster size the job completed in 3,960 seconds; with a 512-byte cluster size the job completed in 7,860 seconds.
- Under NTFS, with a 4K-byte cluster, the job completed in 1,532 seconds; with a 1K-byte cluster size the job completed in 2,963 seconds; with a 512-byte cluster size the job completed in 4,983 seconds.
It's pretty clear that a 4096-byte cluster size provides better performance under both NTFS and FAT-32. More to the point for our investigation of FAT-32 is that it took an hour and a half longer to do the test with a 512-byte cluster size as opposed to a 4096-byte cluster size. That's 41 minutes (4K-byte cluster size) as opposed to 2 hours 11 minutes (512-byte cluster size)...three times longer with the smaller cluster size.
I do want to point out that NTFS with a 4K-byte cluster size performed the best. This echoes back to my previous articles where I insisted that it was well worth it to arrange to have a 4096-byte cluster size on your NTFS boot. In the case of this test suite, using NTFS with 4096-byte clusters results in about one-third the elapsed time as compared with using 512-byte clusters.
In fact, moving from FAT-32 512-byte clusters to NTFS 4K-byte clusters makes the job run in one fifth the time!
I ran a single test on a FAT-16 volume (64K was the only legal cluster size)and we can see that even with a cluster size that large, it still couldn't beat NTFS with 4096-byte clusters.
Now, I know some of you are running huge FAT-32 volumes with 512-byte clusters because of some highly specialized need. In your case, all I can suggest is that you seriously consider using a larger cluster size.
I have also heard protests about "cluster slack" -- that there's wasted space in them-thar big clusters. For them, I can only point out that I paid $129 for a 40GB hard drive a week ago. Suddenly, the last thing on my mind is the disk space I've lost in "cluster slack".
But, for my friend with the 30GB laptop running Windows 2000 only under FAT-32, reinstalling under NTFS with 4K-byte cluster size is a much better idea. Who needs to spend five times longer getting a job done on a laptop?
Obviously, when Diskeeper is working under a file system setup that yields better performance, Diskeeper itself performs better.
So my advice, in general, is stay away from FAT-32 under Windows 2000.
- - - - - - - - - - - - - - - - - - - - - - - - - - -
If you prefer, check the post:
- - - - - - - - - - - - - - - - - - - - - - - - - - -
FAT-32 ON WINDOWS 2000: PENNY-WISE AND POUND-FOOLISH?
by John L. Joseph, Diskeeper Development Section, Executive Software
Editor's Note: If someone is described as "penny-wise and pound-foolish", they are thought to be too careful about very small details and not able to consider larger, more important matters. (Cambridge International Dictionary of English). The expression comes from English (UK) currency, the penny being the much smaller unit and the pound (Sterling) being the larger.
-------------------------------------------------------
As more and more individuals and companies convert to Windows 2000, it's becoming a more and more common question to hear: "Should I use FAT-32?"
It is even more important to answer when I'm told by one user that he just received a desktop machine with Windows 2000 preinstalled on a 20GB drive under the FAT-32 file system, and another user tells me he just received his new laptop with Windows 2000 preinstalled on a 30GB drive under the FAT-32 file system.
Because of other problems I'd seen with the way mail-order computer vendors configured pre-installed systems, I knew I'd have to check it out.
The first place I went to was of course the Microsoft Knowledge Base*.
In article Q100108, I found this discussion:
"Disadvantages of FAT"
Preferably, when using drives or partitions of over 200 MB the FAT file system should not be used. This is because as the size of the volume increases, performance with FAT will quickly decrease."
Wow! That's a bit of a revelation for most computer users I know.
I then found this in article Q154997:
"For most users, FAT32 will have a negligible performance impact. Some programs may see a slight performance gain from FAT32. In other programs, particularly those heavily dependent on large sequential read or write operations, FAT32 may result in a modest performance degradation."
The obvious thing to do was to quantify what "modest" really meant.
I took a 866MHz Pentium III Windows 2000 machine with a DMA-66 IDE hard drive, and carved a 2GB volume out of the hard drive. I then created a program which would place 5200 files of known, predicted size on that volume, filling up the volume till it was 24% free, using all three file systems. I recorded the number of wall-clock seconds it took to complete the job. Here's how it came out:
- Under FAT-16, under which I could only use a 64K-byte cluster size, the job completed in 1,682 seconds.
- Under FAT-32, with a 4K-byte cluster, the job completed in 2,472 seconds; with a 1K-byte cluster size the job completed in 3,960 seconds; with a 512-byte cluster size the job completed in 7,860 seconds.
- Under NTFS, with a 4K-byte cluster, the job completed in 1,532 seconds; with a 1K-byte cluster size the job completed in 2,963 seconds; with a 512-byte cluster size the job completed in 4,983 seconds.
It's pretty clear that a 4096-byte cluster size provides better performance under both NTFS and FAT-32. More to the point for our investigation of FAT-32 is that it took an hour and a half longer to do the test with a 512-byte cluster size as opposed to a 4096-byte cluster size. That's 41 minutes (4K-byte cluster size) as opposed to 2 hours 11 minutes (512-byte cluster size)...three times longer with the smaller cluster size.
I do want to point out that NTFS with a 4K-byte cluster size performed the best. This echoes back to my previous articles where I insisted that it was well worth it to arrange to have a 4096-byte cluster size on your NTFS boot. In the case of this test suite, using NTFS with 4096-byte clusters results in about one-third the elapsed time as compared with using 512-byte clusters.
In fact, moving from FAT-32 512-byte clusters to NTFS 4K-byte clusters makes the job run in one fifth the time!
I ran a single test on a FAT-16 volume (64K was the only legal cluster size)and we can see that even with a cluster size that large, it still couldn't beat NTFS with 4096-byte clusters.
Now, I know some of you are running huge FAT-32 volumes with 512-byte clusters because of some highly specialized need. In your case, all I can suggest is that you seriously consider using a larger cluster size.
I have also heard protests about "cluster slack" -- that there's wasted space in them-thar big clusters. For them, I can only point out that I paid $129 for a 40GB hard drive a week ago. Suddenly, the last thing on my mind is the disk space I've lost in "cluster slack".
But, for my friend with the 30GB laptop running Windows 2000 only under FAT-32, reinstalling under NTFS with 4K-byte cluster size is a much better idea. Who needs to spend five times longer getting a job done on a laptop?
Obviously, when Diskeeper is working under a file system setup that yields better performance, Diskeeper itself performs better.
So my advice, in general, is stay away from FAT-32 under Windows 2000.
- - - - - - - - - - - - - - - - - - - - - - - - - - -
If you prefer, check the post:
danielshoes
Senior member
I have 98 and 2000 on my computer-- 98 on drive C and 2000 on drive D. Both are running FAT32 and 2000 crashed big time. I was able to run in 98 and go into D drive into 2000's folders and move the files I wanted to save before I reinstalled 2000. I was able to save everything. Could I have done that if D had been NTFS?
JackBurton
Lifer
<< . I was able to save everything. Could I have done that if D had been NTFS? >>
YES you could but you'd have to go about a different way to pull your data off 2000. Install that drive on another PC running 2000 and set it up as slave. After that just pull your data off. Easy. 🙂
Just don't encrypt your drive because then your SOL. 😉
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 25K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 24K
-
-
