I have been searching for one of these online and cannot seem to find anyone selling them.

Can anyone point me to a relaible reseller??

Thanks

JAA
Austin TX

http://www.pricegrabber.com/search_getprod.php?masterid=4347730&search=MK4004GAH

http://www.pcconnection.com/ProductDetail?sku=5191416

http://www.psism.com/hdd1524.htm

Thanks. Ordered and on the way (I hope!!)

JAA

just wondering, which one would give a better performance IF both HD exist in the market? 60 GB 4200 rpm or 40 GB 5400 rpm ?  i might keep my TR for a while until i see sony comes up with new latop with camera just like TR, in that case i need a better HD to take advantage of the processor coz now the 40 GB 4200 rpm is slowing the whole operation

thanks.

[quote author=“gameson”]the 40 GB 4200 rpm is slowing the whole operation

thanks.

Agreed. Our main problem is the slow hd.

i disagree…

top 5 reasons for a system to run slow

1) too much shit into the memory

2) out of memory

3) Fragmented Hard Drive

4) Faulty Mobo/Memory/HDD

5) Compatibility issues.

The drive in the TR’s T’s whatever are fast enough. They are proportional for the system ... i.e. memory/video/cpu.

If you need a faster drive, it is so easy to do it with the TR’s its not even funny.

get an external usb 2.0 enclosure with sata interface

buy a 10,000RPM 8mb cache WD raptor 36 or 73 gigs.

install windows on that and be done..

also if you dont want to go through that trouble, a very easy fix will speed up your machine, just move the paging file to a second partition.

I assume everyone splitted the disk into at least 2 partitions..

S

[quote author=“DayWalker”]I assume everyone splitted the disk into at least 2 partitions..

S

Bad assumption, I hate partitions. My PCs don’t have one partitioned drive (besides the main one).

Erik

if u split it and move the paging file, u will notice difference.

the reasons are

1) the paging file will be at a drive which is less used

2) fat table on that drive will be smaller and earily accesable

3) less files that allows faster access

by design NTFS sux, so this is a little tweak around it…

[quote author=“DayWalker”]
get an external usb 2.0 enclosure with sata interface

Do it with Firewire. USB 2.0 uses up CPU time while transfering data, and the last thing you want is saturating your CPU by just accessing your hard drive.

buy a 10,000RPM 8mb cache WD raptor 36 or 73 gigs.

The raptor is not superfast because of its transfer rate, it’s because of its incredibly low access times. And you won’t benefit from this ny using a USB2.0 or Firewire interface. No sir, not gonna happen.

1) the paging file will be at a drive which is less used

If it’s on another drive: yes. If it’s on another partition: no. The head still has to move. There’s just no difference. Wether the operating system is accessing logical block 2344 or 2136254, speed is the same (besides from the disk zones - but those have to do with diameter and platters, not with partitions).

The benefit with using a paging file on another physical drive is that both drives, the one the computer’s working on and the one where the paging file resides, can be accessed independently and without heavily moving the read/write-heads around.

2) fat table on that drive will be smaller and earily accesable

A smaller FAT (it’s the acronym for ‘file access table’, so no need to say ‘FAT table’) doesn’t give you any speed increase. why should it?
yes, okay, checking the filesystem (scandisk) or the initial mount at boot may be a tad faster… milliseconds that occur only once during the PCs uptime. Not really what I would call a priority.

3) less files that allows faster access

Again: why?

by design NTFS sux, so this is a little tweak around it…

Well, NTFS certainly isn’t the pointiest needle in the box, but it has its benefits.

[quote author=“DayWalker”]
The drive in the TR’s T’s whatever are fast enough. They are proportional for the system ... i.e. memory/video/cpu.

Just let me add that there isn’t any proportionality between CPU/memory and harddisk (and certainly not between Video and harddisk).

The harddisk is thousands of thousands times slower than the CPU and the memory, always. As long as the harddrive is slower than the memory (and you can’t even imagine how astronomically f*ing big that difference is) a faster harddrive will always give you a speed improvement when accessing it.

The CPU has nothing to do with it, even if it would be - theoretically - too slow. DMA is the magic word, direct memory access. The harddrive doesn’t push data through the CPU into the memory. The ATA bus pushes it directly into the memory where it’s available for use by the CPU or other devices. That’s why I said you wouldn’t want USB 2.0 for harddrives - it’s not as good in this aspect as firewire. Not because it doesn’t implement DMA, the data still goes directly into memory, but because there’s a lot of USB protocol overhead which the CPU has to handle.

[quote author=“FrodoID”]

1) the paging file will be at a drive which is less used

If it’s on another drive: yes. If it’s on another partition: no

2) fat table on that drive will be smaller and earily accesable

A smaller FAT (it’s the acronym for ‘file access table’, so no need to say ‘FAT table’) doesn’t give you any speed increase. why should it?

3) less files that allows faster access

Again: why?

by design NTFS sux, so this is a little tweak around it…

Well, NTFS certainly isn’t the pointiest needle in the box, but it has its benefits.

Agreed with most of the points… what kills drive speed in most of the TRs (depending on the use) is seek time and having the swap in a separate partition will not improve that… actually, it will make it worst.  Furthermore, since modifying the swap location shows change in performance, it basically means that it is heavily used… ergo… get more memory   Fragmentation will definitely make it slower and having an improperly maintained filesystem with a permanently changing swap will create lots of fragmentation.

Also… FAT (which means “file allocation table”) sizes are not defined by the number of files currently in the disk.  “Smaller” FATs are not related to more partitions but rather to the version of the OS and the geometry of the disk.  Additionally, NTFS does not use FATs to begin with, buth rather MFTs that are a different animal.  $MFTs are basically relational databases of file attributes and they DO grow with the addition of files and journaling entries (but moving ONE - swap-  file to another partition will not make a difference)

In any case, to each its own… and although NTFS is not perfect (even NTFS5), it is a qualitative leap from FAT incarnations and even from EXT2 (which does not even support journaling).  Why do you think that the linux community came out with EXT3?

Just my point of view… YMMV

Regards,
Marcelo

Anyways, can a USB 2.0 handle the data flow from a 10,000 RPM HD? I mean, is 480 Mbps enough to support a 10,000 RPM hd data flow in its full speed? Waht is its Mbps output?

[quote author=“mfabre”]
Also… FAT (which means “file allocation table”)

Argh… I swear it was a typo

sizes are not defined by the number of files currently in the disk.  “Smaller” FATs are not related to more partitions but rather to the version of the OS and the geometry of the disk.

Yes. We were talking about partitioning. With two small partitions you have two smaller FATs, compared to one big partition with one big FAT. So DayWalker was right, partitioning gives you smaller FATs, but it doesn’t help. (At least that’s how I understood him. But maybe he really talked about smaller FATs by having less files, which is of course nonsense.)

Additionally, NTFS does not use FATs to begin with, buth rather MFTs that are a different animal.

Yes, DayWalker was saying that NTFS “sucks by design” so he wouldn’t use it and we were talking about FAT only. Possibly the reason why he wants partitioning, to overcome the whatever-it-is-now size limit of FAT. (Again, that’s what I understood. There’s still the possibility that he meant FATs on NTFS, which would also be wrong.)

$MFTs are basically relational databases of file attributes and they DO grow with the addition of files and journaling entries (but moving ONE - swap-  file to another partition will not make a difference)

I think NTFS is a pretty impressive beast, what with all those several streams per file (if you can still call them files) and extended attributes. You’re completely right, it is more of a relational database than a traditional filesystem…

One question: does adding more files to the filesystem with NTFS really make a change in performance, just because of the additional records? Shouldn’t a strong hash based index or some kind of ordered binary tree referencing the records make sure that that won’t happen too fast?

In any case, to each its own… and although NTFS is not perfect (even NTFS5), it is a qualitative leap from FAT incarnations and even from EXT2 (which does not even support journaling).  Why do you think that the linux community came out with EXT3?

Well, ext3 is basically just ext2 with journalling. They could also have kept the name ext2 with just adding the journaling, but they also wanted to start over with their code base, making it cleaner. To keep both code bases, the stable ext2 code base and the more experimental but cleaner ext3 code base, they chose to have two different names. Although I’m using it almost exclusively, I really don’t think ext2/ext3 is the best filesystem - it’s just the one that I currently trust the most.

Do you know of the plans for the next ReiserFS generation? Really impressive and this will punch NTFS straight out of the ring. ReiserFS will also implement multiple streams, replacing those damn extended attributes, and introduces the new “every file is a directory” concept. A pretty controversial thing which much discussion, because it of course isn’t just a ReiserFS thing but it will change the whole filesystem semantics. If you can chdir into every file, that really is a hell of a semantic change

However, Linus is pretty positive about it and I’m also a little bit excited about how far it will get. The Linux Kernel Mailing List has some nice examples about what can be done with it within the discussions, and there really is some awesome stuff which you wouldn’t think of in first place.

Along with that, the hashing and directory tree algorithms are already superior to NTFS.

Funny, I never used ReiserFS at all. Maybe I should finally consider using it.

Just my point of view… YMMV

I totally agree with all of your points

Regards,
Julien

[quote author=“Lourenco”]Anyways, can a USB 2.0 handle the data flow from a 10,000 RPM HD? I mean, is 480 Mbps enough to support a 10,000 RPM hd data flow in its full speed? What is its Mbps output?

Keep in mind that the 10000rpm aren’t necessarily very important for the data throughput, but a big thing for the access time.

Other than that: a WD Raptor has normally around 55MB/s linear constant data throughput. I just tried it out with my Silicon Imaging 3112. That’s 440MBps - theoretically. There’s still some overhead to count in. So it’s pretty close to the limit, but it’s possible. Again, theoretically.

The bigger problems are latency time (which increases your access times, which are in most of your usage scenarios more important than data throughput) and CPU usage for USB overhead handling, which means that your CPU is actually busy while transferring data from and to the harddisk and you just don’t want that - it’s not necessary and especially poisonous for multitasking.

Regards,
Julien