Slashdot Mirror

"* *Actual* performance problems due to fragmentation - outside of a few corner cases - are basically nonexistant. "

Yep. That's why I have to run defragmenter on our build server every week...

Also, Windows is notoriously slow with file operations. It's not directly related to NTFS, but more to extremely inefficient VFS stack.

"* Can you explain what you mean by "it's done above the VFS layer" ? Surely you're not trying to argue symlinks and shortcuts are the same thing ? "

http://neosmart.net/blog/2006/vista-symlinks-revisited/

"* RAID is handled at the block device level, not the filesystem level (and many, many people believe putting RAID into the "filesystem" is an architecturally bad thing, so that's hardly something it can be plainly criticised for)."

However, filesystem-level RAIDs have a lot more functionality than block-level RAIDs. Look at ZFS or BTRFS.

"* Do you have a source for up-to-date benchmarks ?"

I have my own set of benchmarks. Well, NTFS on Windows is almost always slower (and quite often like 100 _times_ slower) than Linux filesystems.

http://rsdn.ru/File/37054/benchmark.zip - this is the source.

http://rsdn.ru/forum/philosophy/1710544.1.aspx - this is a post with benchmark results (in Russian, sorry - I can translate if you have any questions)

http://rsdn.ru/forum/philosophy/1712431.aspx - this post contains this benchmark, slightly adapted.

I regularly re-run these tests. So far, Windows is only getting slower compared to Linux.

I've recently created a multithreaded version of this test. Well, let's say that NTFS sucks so badly, that it's hard to understand how MS has managed to achieve this.

"* *Actual* performance problems due to fragmentation - outside of a few corner cases - are basically nonexistant. "

Yep. That's why I have to run defragmenter on our build server every week...

Also, Windows is notoriously slow with file operations. It's not directly related to NTFS, but more to extremely inefficient VFS stack.

"* Can you explain what you mean by "it's done above the VFS layer" ? Surely you're not trying to argue symlinks and shortcuts are the same thing ? "

http://neosmart.net/blog/2006/vista-symlinks-revisited/

"* RAID is handled at the block device level, not the filesystem level (and many, many people believe putting RAID into the "filesystem" is an architecturally bad thing, so that's hardly something it can be plainly criticised for)."

However, filesystem-level RAIDs have a lot more functionality than block-level RAIDs. Look at ZFS or BTRFS.

"* Do you have a source for up-to-date benchmarks ?"

I have my own set of benchmarks. Well, NTFS on Windows is almost always slower (and quite often like 100 _times_ slower) than Linux filesystems.

http://rsdn.ru/File/37054/benchmark.zip - this is the source.

http://rsdn.ru/forum/philosophy/1710544.1.aspx - this is a post with benchmark results (in Russian, sorry - I can translate if you have any questions)

http://rsdn.ru/forum/philosophy/1712431.aspx - this post contains this benchmark, slightly adapted.

I regularly re-run these tests. So far, Windows is only getting slower compared to Linux.

I've recently created a multithreaded version of this test. Well, let's say that NTFS sucks so badly, that it's hard to understand how MS has managed to achieve this.

"* *Actual* performance problems due to fragmentation - outside of a few corner cases - are basically nonexistant. "

Yep. That's why I have to run defragmenter on our build server every week...

Also, Windows is notoriously slow with file operations. It's not directly related to NTFS, but more to extremely inefficient VFS stack.

"* Can you explain what you mean by "it's done above the VFS layer" ? Surely you're not trying to argue symlinks and shortcuts are the same thing ? "

http://neosmart.net/blog/2006/vista-symlinks-revisited/

"* RAID is handled at the block device level, not the filesystem level (and many, many people believe putting RAID into the "filesystem" is an architecturally bad thing, so that's hardly something it can be plainly criticised for)."

However, filesystem-level RAIDs have a lot more functionality than block-level RAIDs. Look at ZFS or BTRFS.

"* Do you have a source for up-to-date benchmarks ?"

I have my own set of benchmarks. Well, NTFS on Windows is almost always slower (and quite often like 100 _times_ slower) than Linux filesystems.

http://rsdn.ru/File/37054/benchmark.zip - this is the source.

http://rsdn.ru/forum/philosophy/1710544.1.aspx - this is a post with benchmark results (in Russian, sorry - I can translate if you have any questions)

http://rsdn.ru/forum/philosophy/1712431.aspx - this post contains this benchmark, slightly adapted.

I regularly re-run these tests. So far, Windows is only getting slower compared to Linux.

I've recently created a multithreaded version of this test. Well, let's say that NTFS sucks so badly, that it's hard to understand how MS has managed to achieve this.

Technically, it's not that hard for MS to add this support. Objective-C/C++ is just a plain C with some preprocessor magic, and VS in its current iteration is quite extensible.

Certainly, adding support for another language is possible as a simple add-on (for example, Nemerle language has one - http://rsdn.ru/article/nemerle/Nemerle.VsIntegration-en.xml ).

I've read (Russian) that the following code crashes in runtime:

let comp x y = x > y
let result = comp (new Dictionary<object, object>) (new Dictionary<object, object>)

(I am not an F# expert, I am a Haskell user;)

Java code is usually larger than code in other languages, true.

But currently Java has the best tools in existence. IntelliJ IDEA is the best IDE out there.

In fact, power of Java tools is so great, it can argued that Java+IDE is in fact a higher-level language. For example, C# has LINQ for statically-typed queries checked at compile time, Java has nothing of that sort in the core language. However, "edit-time" inspections in IDEA provide even better support for checking queries: http://files.rsdn.ru/37054/HQLBug.png (here a non-existing field and incorrect placeholders are highlighted).

Another example, Java doesn't have closures, but the can be effectively added by IDE: http://blogs.jetbrains.com/idea/2009/03/closure-folding-in-intellij-idea-9-maia/

So writing Java in a good IDE for me is better than writing code in a great language in a plain text-editor (i.e. ViM, Emacs, TextMate).

http://www.rsdn.ru/forum/philosophy/1710544.1.aspx - sorry, it's in Russian. You can download benchmark here: http://www.rsdn.ru/File/37054/benchmark.zip Basically, it creates, stat()s and deletes lots of files. As you can see, performance in Windows is quite poor.

I have several more microbenchmarks and _all_ of them work faster on Linux. As a not-very-micro-benchmark: git works way faster on Linux.

And it's not the problem of NTFS itself, because ntfs-3g on my computer _still_ works faster for a lot of operations than the native NTFS in Windows!

http://www.rsdn.ru/forum/philosophy/1710544.1.aspx - sorry, it's in Russian. You can download benchmark here: http://www.rsdn.ru/File/37054/benchmark.zip Basically, it creates, stat()s and deletes lots of files. As you can see, performance in Windows is quite poor.

I have several more microbenchmarks and _all_ of them work faster on Linux. As a not-very-micro-benchmark: git works way faster on Linux.

And it's not the problem of NTFS itself, because ntfs-3g on my computer _still_ works faster for a lot of operations than the native NTFS in Windows!

You can check my posting history. I commented on a lots of other articles concerning Russia or Russian language.

Or here's my profile on a Russian-speaking site (Russian Software Developers Network): http://rsdn.ru/Users/37054.aspx , where I have about 18000 messages.

For those who can read Russian here's a discussion about [X]Emacs here: http://rsdn.ru/Forum/?mid=1322168

PS: I'm currently developing my own IDE (called FlexIDE) because _none_ of UNIX IDEs (and this includes Emacs and ViM) is good enough for me...