Hi! We see you’re using an ad-blocker. We’re fine with that and won’t stop you visiting the site.
But as we’re losing ad-revenue from this then why not make a donation towards website running costs?. Or you could disable your ad-blocker for this site. We think you’ll find our adverts are not overbearing!
Posted: Thu Jan 15, 2004 10:26 am Post subject: Is fragmentation an issue on storage cards?
Hi there,
I have seen several posts that refer to defrag utilities for CF and SD cards.
Is there any benefit in using these?
I know that standard hard drives (so presumably micro drives) benefit from unfragmented files because of the reduced amount of physical head travel required for read/write seeks. But seeing as storage cards have no mechanical moving parts, is fragmentation an issue?
Hmm...I can't recall ever seeing a system RAM defragger.
So...Are storage card defrag utilities a genuinely useful asset or just a cynical money maker that takes advantage of peoples fear and ignorance?
Some storage cards do have moving parts because they are mini hard drives. But, just because a storage card isn't a hard drive, don't think it doesn't get fragmented. Fragmentation is all about how the card is formatted, and how that type of format utilises and writes data. Storage cards use the same file systems as seen under Dos, Windows 3.1, Windows 95, Windows 98 and Windows ME, so you will still get fragmentation.
A system RAM defragger is basically called hitting the soft reset button. There's no call for one.
I appreciate that storage cards use standard file systems and that fragmentation does occur in these file systems. My question was more to do with the different hardware implementations and the way that data is accessed on solid state devices compared to disk drives and whether or not fragmentation is actually an issue on flash cards.
On disk drives fragmentation impedes performance because there's a variable seek time, depending on the distance the head needs to move, and how much the disk needs to rotate before the required sector can be accessed. Data access is also faster when the sectors occupy outer edges of the disk. Disks are faster when the sectors are arranged without fragmentation to minimize the seek time, and a fragmented disk is slow because the head spends a lot of time travelling back and forth all over the place.
In contrast, flash memory access is completely independent of the actual sector address. There is no head to move back and forth. The addressing is all done electronically, so accessing different sectors on a flash card would take the same amount of time regardless of the actual location on the card. It doesn't matter whether the sectors of a particular file are contiguous or not.
I'm still not convinced that fragmentation on a flash card is an issue.
My question was more to do with the different hardware implementations and the way that data is accessed on solid state devices compared to disk drives and whether or not fragmentation is actually an issue on flash cards.
Mick, no difference! Fragmentation is fragmentation. If a 1mb file is fragmented 20 times into 51k blocks, then it's fragmented and it means that no matter what storage device you have, reading that file in fragmented blocks will be slower than reading it in whole. The only difference comparing a memory chip storage card to a hard drive is that the speed will be worse on a hard drive due to access times, but you will still see speed issues on memory chip storage cards if they are heavily fragmented.
MickH wrote:
In contrast, flash memory access is completely independent of the actual sector address. There is no head to move back and forth. The addressing is all done electronically, so accessing different sectors on a flash card would take the same amount of time regardless of the actual location on the card.
Yes, the read time will still be the same for each sector, but don't forget there's extra time for reading 20 times the data from the FAT table on where each sector is stored, and then having to go and read the real data. Sure it won't be as slow as a hard drive when it's heavily fragmented, but there is still fragmentation and there still can be speed issues, especially if you're trying to load a 200mb map file that's fragmented into 50 different storage blocks on a storage card.
Another thing to bear in mind is look at the speed of SD cards compared to Compact Flash, CF are much faster, so you would also see a significant difference on a fragmented SD card compared to a fragmented CF card due to access times on SD cards.
My understanding was that each entry in the FAT table contained a reference to the next required sector (i.e. a linked list), until an EndOfFile marker was found, so when accessing files I presumed that once the initial sector of the file had been determined via the file manager, the IO routine would traverse through the FAT creating a list of the required sectors. So accessing a file occupying sectors {895, 2045, 6032, 458} would take the same time as accessing a file in (895, 896, 897,898}.
I didn't realise that whole sections of the FAT were read at a time.
Thanks for the info.
I'll check out the SoftWinter app.
Joined: Nov 24, 2003 Posts: 1441 Location: Swansea
Posted: Fri Jan 16, 2004 8:25 pm Post subject:
I found that if I put the storage card in a card reader connected to my PC, my pc's defrag program (Executive software Diskeeper) will defrag the card in a few seconds.
Robin
Joined: 17/05/2003 02:26:21 Posts: 3747 Location: Bedfordshire, UK
Posted: Mon Jan 19, 2004 1:07 am Post subject:
I find Diskeeper doesn't behave that sensibly on storage cards, particularly those with not much full space. There again, I last tried with Diskeeper 7.0 build 430. I've upgraded to Diskeeper 8.0 Professional yesterday (and installed the patch to move to the very the latest 8.0 build), so I could try again and see if version 8 behaves any more sensibly.
I recommend Diskeeper highly for hard disks. It is certainly a fallacy that NTFS partitions aren't affected by fragmentation. I believe my investment in a very high end HD system in my workstation would be wasted without it. It seems silly to go to the expense of a PCI-X U320 SCSI controller and two Seagate Cheetah 15K.3 U320 SCSI hard disks, then refuse to spend 30 pounds or so for efficient automatic defragmentation. That said, Diskeeper makes just as much sense for people with regular hard disks in their computers.
Just make sure that you buy the right edition of Diskeeper 8.0. With 7.0, if you weren't on a system where Diskeeper was remote controlled from an administrative setup, you could install the cheap Diskeeper Home edition. With 8.0, that isn't always the case - if you use Windows XP Professional, you need Diskeeper Professional (which replaces the old Diskeeper Workstation edition).
Like Dave, I have SoftWinter's StorageTools on my iPAQ. However, though I haven't yet really played with the defragmentation utility, I have a copy of Anton Tomov's Pocket Mechanic - and that looks, if anything, to be a better buy - it does more for similar money. However, StorageTools is a solid offering and I wouldn't want to dissuade anyone from using it.
Storage card performance is definitely hit by fragmentation - comparisons to RAM are invalid, as the read speed of a storage card is much lower than the read speed of RAM.
Joined: 24/06/2003 00:22:12 Posts: 2946 Location: Escaped to the Antipodies! 36.83°S 174.75°E
Posted: Mon Jan 19, 2004 1:26 pm Post subject:
DavidW wrote:
Storage card performance is definitely hit by fragmentation - comparisons to RAM are invalid, as the read speed of a storage card is much lower than the read speed of RAM.
I think comparisons with RAM are valid because the read speed is not relevant here - it's the seek time that causes a fragmented file to be read more slowly since the hard disk has to do multiple seeks to read the file.
If the seek time is insignificant and data can be read just as quickly from any part of the media (as it is in a solid state memory card) , then how can the card be affected by fragmentation?
Joined: 17/05/2003 02:26:21 Posts: 3747 Location: Bedfordshire, UK
Posted: Mon Jan 19, 2004 10:58 pm Post subject:
Skippy wrote:
DavidW wrote:
Storage card performance is definitely hit by fragmentation - comparisons to RAM are invalid, as the read speed of a storage card is much lower than the read speed of RAM.
I think comparisons with RAM are valid because the read speed is not relevant here - it's the seek time that causes a fragmented file to be read more slowly since the hard disk has to do multiple seeks to read the file.
If the seek time is insignificant and data can be read just as quickly from any part of the media (as it is in a solid state memory card) , then how can the card be affected by fragmentation?
My apologies for not replying that accurately and glossing simply by saying that storage cards are slower than RAM. I was trying to reply to a lot of posts quickly.
The seek time is significant on storage cards. The benchmark program I've got to hand (Anton Tomov's Pocket PC Mark) allows me to do some quick tests. On my "Made in Japan" 256MB Lexar SD card, which is one of the better performing types out there, using an iPAQ 3970 with ROM 4.00.08 (Pocket PC 2003 - officially released ROM), the seek score is 14529.98 positions/sec (unfortunately positions aren't defined - I suspect that could be sectors or clusters). To try to make the results more accurate, I reformatted the storage card first and carried out the tests with the card blank.
However, truly accurate benchmarking (particularly for something like seek time) is always a little tricky on the Pocket PC. It's also the case that different combinations of storage card, Pocket PC hardware and storage card drivers perform differently.
It is an observable phenomenon that, particularly on lesser storage cards (such as Sandisk SD cards), fragmentation of the map files can severely impact performance of navigation software.
Meanwhile, fragmentation in the internal RAM filing system can be ignored - being RAM, seek times are essentially meaningless.
I'm sorry that this answer isn't that technically rigorous - much of the problem is that full details of the SecureDigital protocol are proprietary and are only available after entering into an agreement. Architectural details of various company's SD cards are, of course, company confidential as are implementation details of various company's SD drivers.
Posted: Today Post subject: Pocket GPS Advertising
We see you’re using an ad-blocker. We’re fine with that and won’t stop you visiting the site.
Have you considered making a donation towards website running costs?. Or you could disable your ad-blocker for this site. We think you’ll find our adverts are not overbearing!
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
Or you could disable your ad-blocker for this site. We think you’ll find our adverts are not overbearing!
Hi! We see you’re using an ad-blocker. We’re fine with that and won’t stop you visiting the site.
But as we’re losing ad-revenue from this then why not make a donation towards website running costs?. Or you could disable your ad-blocker for this site. We think you’ll find our adverts are not overbearing!