Windows 7

[Sxotty]

I don't know, it looks like some people had to take steps in addition to the ones from the original post and the solution for Windows 7 in the final post involves having two PC's at your disposal, but here's the link if you're feeling adventurous. ...

Hey thanks man. I already played through it with virtual pc with XP installed and it worked perfect. I found similar threads,but the links were always dead to download the patch. I guess I shoulda just come here first

 

[spacemonkey]

Whatever happened to that new database file system which was supposed to replace NTFS, but was axed in Vista?

 

[Scali]

Whatever happened to that new database file system which was supposed to replace NTFS, but was axed in Vista?

I suppose it got reinvented as Windows Search + the new 'Libraries' system in Windows 7.
Afaik it was never going to replace NTFS anyway, it was just a database running on top of NTFS.

 

[Kaotik]

Yep, it was database "addon" for NTFS, and it still might come though Scali might be right at least to some extent with libraries, there has been some rumors about the internal RC2 version having some new WinFS stuff in it, or early version of WinFS

 

[ShaidarHaran]

It's called WinFS and we've only been hearing about it for 10 years now.

 

[suryad]

All I want MS to do is switch to ext4 or something equivalent to that. Heck ReiserFS would be perfect but that is not going to happen. I just think it is ludicrous that we still need to be stuck with a filesystem that needs to be defragged.

That said I am running Windows 7 RC as my main OS for the past few days and I like it very much. I am actually surprised to see every app running a lot smoother than they did in XP x64 and that includes games! I like it quite a bit actually. I am definitely going to grab a copy of W7 once March 2010 rolls around.

 

[ninelven]

Sadly, I doubt we will see anything on this front before the next release. Hopefully, Btrfs will come along and MS will embrace it.

Still, Windows 7 is very solid. It just works and does what you want it to do, which more than makes up for NTFS IMO.

 

[suryad]

Sadly, I doubt we will see anything on this front before the next release. Hopefully, Btrfs will come along and MS will embrace it.

Still, Windows 7 is very solid. It just works and does what you want it to do, which more than makes up for NTFS IMO.

Now if Creative could sort their audio drivers out that would make it perfect for me Their betas are pretty good but the 7.1 does not work quite well with DVD playback.

 

[Davros]

suryad, win7 and eax ?

 

[pcchen]

All I want MS to do is switch to ext4 or something equivalent to that. Heck ReiserFS would be perfect but that is not going to happen. I just think it is ludicrous that we still need to be stuck with a filesystem that needs to be defragged.

Any file system will eventually require defragment, and that includes ext4 (which has an online defragment tool planned).

 

[suryad]

suryad, win7 and eax ?

I am pretty new to Win 7 so forgive me if I dont understand what you were trying to ask....

 

[Scali]

All I want MS to do is switch to ext4 or something equivalent to that. Heck ReiserFS would be perfect but that is not going to happen. I just think it is ludicrous that we still need to be stuck with a filesystem that needs to be defragged.

Fragmentation is a variation of the knapsack problem, which is an NP-complete problem.
Now unless someone figured out a way to solve NP-complete problems in realtime, so they can be used in a filesystem, I don't think there is any chance of a filesystem that doesn't fragment.
I think people mistake the lack of proper defragmentation tools on *nix for these systems not being plagued by the problem of fragmentation.

 

[Davros]

does you get eax in supported games under win7 or do you need a program like creative alchemy

 

[Bludd]

does you get eax in supported games under win7 or do you need a program like creative alchemy

Same as Vista. Maybe ALchemy works with Win7 but OpenAL is the way to go anyway.

 

[suryad]

Fragmentation is a variation of the knapsack problem, which is an NP-complete problem.
Now unless someone figured out a way to solve NP-complete problems in realtime, so they can be used in a filesystem, I don't think there is any chance of a filesystem that doesn't fragment.
I think people mistake the lack of proper defragmentation tools on *nix for these systems not being plagued by the problem of fragmentation.

Hehe interesting. I have to read up on the knapsack problem. You know what the worst part is? I actually am a CS grad LOL. What strikes me as odd is that with Windows, you can gradually notice a degradation of performance over time as the file system gets more and more fragmented. However using Linux at work and using various tools when doing builds and such, I notice the times are always the same yet in Windows it gets slower and slower and slower over time, till one day I cannot stand it anymore and I make it a point to defrag over the weekend. I wonder if disk fragmentation is not the cause of that drop in performance over time then what is?

does you get eax in supported games under win7 or do you need a program like creative alchemy

Well to be honest the only game I have installed right now is The Last Remnant and surround sound does work on it. The only thing is that the bass is a bit subdued. All I remember in XP x64 is that the sound effects had a lot more oomph to it. And once again I have no idea what Creative Alchemy is and OpenAL.

 

[Scali]

Hehe interesting. I have to read up on the knapsack problem. You know what the worst part is? I actually am a CS grad LOL.

Ouch

What strikes me as odd is that with Windows, you can gradually notice a degradation of performance over time as the file system gets more and more fragmented. However using Linux at work and using various tools when doing builds and such, I notice the times are always the same yet in Windows it gets slower and slower and slower over time, till one day I cannot stand it anymore and I make it a point to defrag over the weekend. I wonder if disk fragmentation is not the cause of that drop in performance over time then what is?

Simple way to verify:
Defragment the disk, then see if the performance is back. If not, fragmentation was not the problem.

My personal experience with FreeBSD is that performance is so utterly horrible compared to Windows to begin with, that fragmentation is not really a factor in the total performance.
Fragmentation DOES happen on UFS (the fsck tools can report the fragmentation ratio), you just can't do anything about it.
Thing is just, I have this box with a disk that is capable of about 60 mb/s in Windows, but with FreeBSD I get about 10 mb/s.

Different filesystems may have different strategies for allocation, but it's just impossible to have a fragmentation-free system. Fragmentation will ALWAYS happen when you delete one file, then re-use that space for a new file of different size.
But actually it's not the filesystem itself, but the allocation strategy. It's perfectly possible to continue using NTFS, but change the allocation strategy, so you will get a different fragmentation profile. I wouldn't be surprised if Microsoft tweaks this strategy every now and then, so eg XP may have way different fragmentation issues than Windows 7, even though they both use NTFS, and can access eachothers partitions.

 

[suryad]

Hey Scali,

Well when I say CS grad I meant CS undergrad hehe minor correction there. Kinda weird but I never really enjoyed the subject yet I am in the software development field. I hate theory, I would rather do stuff by learning but ah well.

Anyway, I have done the test before and defragmentation was the problem. For example my build was taking close to 3 min for a project. Do the defrag (offline defrag first, then online) and it drops to as quick as 1 min 8 sec which is the average linux build time.

I see what you are saying but it is hard to believe that Linux (Ubuntu's latest distro) is masking the problem of fragmentation through some smart algorithm or something like that.

Did some reading and I found this very nice example. http://geekblog.oneandoneis2.org/index.php/2006/08/17/why_doesn_t_linux_need_defragmenting

Seems to me that if yo uare a bit smart, one can go ahead and prevent fragmentation in a lot of scenarios! Wonder why MS hasnt done that yet.

 

[Scali]

Seems to me that if yo uare a bit smart, one can go ahead and prevent fragmentation in a lot of scenarios! Wonder why MS hasnt done that yet.

I wonder why people assume that MS isn't doing this.
I'm quite sure that they do. It's rather trivial. Thing is, it's not a catch-all. Perhaps the problem is more related to the amount of disk activity and all, and how applications tend to use the filesystem.

One nice example is with file transfer software (eg p2p clients). Some naive software will just write each packet of data to the downloading file as it is received. Especially if you're downloading multiple files at the time, the packets will be written in a rather random order.
A smarter application will pre-allocate a complete file before it starts downloading, and then you have a nice contiguous reserved space where all the packets will fall in place as they are received.

Having said all that, I never had any performance problems with compiling. And I even use file compression. Source code is all ASCII anyway, so it compresses very well.

And if you read through the replies on that article, you'll see that various people there say that linux does fragment aswell, this article actually proves this indirectly.

 

[suryad]

Its not that I have problems compiling...its just that it takes longer and longer over time the more fragmented the drive gets. This is in XP 32 bit by the way...it could well be that MS has updated the filesystem in the mean time between XP 32 and Windows 7. I have not yet tried to do any work on my W7 machine...so maybe that problem wont be around anymore...which would totally rock.

 

[Davros]

Thing is just, I have this box with a disk that is capable of about 60 mb/s in Windows, but with FreeBSD I get about 10 mb/s.

could that be a raid driver problem, my p4 machine gets about 5mb/sec without the via raid driver installed, with it installed i get about 60 mb/sec even though i dont use raid (single sata disk)