Naked Science Forum
Non Life Sciences => Geek Speak => Topic started by: FuzzyUK on 11/06/2011 09:13:26
-
I bought an 8Gb SDHC memory card recently but it only indicates 7.2Gb as free. That's 10% less than stated; 800Mb seems to be lost in thin air. Anyone with an explanation please?
-
Some capacity is taken up by info related to the filesystem placed on the card, and some is lost in salesmanship - the manufacturer says 8G is 8 times 1000 megabytes, and 1M is 1000 kilobytes, 1k is 1000 bytes. The computer says 8G is 8 times 1024 Megabytes, and 1M is 1024 kilobytes, whilst 1k is 1024 bytes.
It is a normal thing, driven by sales and marketing fluffheads who do not understand that computers actually work in base 2 arithmetic, not base 10.
-
Do you have the card in a computer?
What is the "total" capacity on the card?
Do you have your computer enabled to view hidden & system files?
Sometimes companies will add utilities or something that you don't really need. If it happens to be bootable when you don't need (or want) it to be bootable, that can waste a lot of space, and could potentially be dangerous.
Hard drives used to be sold with an unformatted capacity. Once formatted, a certain amount of space would be lost. So, for example, an 80MB hard drive would be formatted to about 72MB, but I thought that practice had largely been stopped.
8GB is 8x1024x1024x1024 bytes = 8,589,934,592 bytes. However anything less would be incorrect, but it is possible that you were sold a card that was 8,000,000,000 bytes which would turn out to be 7.45 GB.
-
I checked the new card in a USB adaptor before I used on it by means of an iMac. It indicated FAT32 formatting and that 256kb was already used, so I assumed that would be the amount deducted off the full 8Gb capacity.
It seems this deception and has probably been going on for several years. I'm surprised Trading Standards have not jumped on the band wagon by now and forced the industry to come clean with their marketing ploys.
Checking on Toshiba's website (although my card was made by PNY), I see they state their particular SDHC Memory Card card capacity as 8Gb. Another column says: Minimum consumer-usable memory capacity*1 Approx. 7.2GB
*1 then tells us this:
The Product is labeled based on the memory chip (s) it contains, not the amount of memory capacity available for data storage by the end user. Part of the capacity is reserved for card functionality. Above list is Minimum consumer-usable memory capacity is approximately.(For purposes of measuring memory capacity in this context, 1 GB = 1,073,741,824 bytes).
Which equates to jargon for "we win - you lose".
-
Yes - I was just about to post but you beat me to it.
The chip is 8GB, but some of it is reserved for internal functions, and some of it may not actually work [:D]. They map out the "bad sectors" when they test the chip during the manufacturing process. That's why they are a bit vague about how much space the user actually gets.
The numbers game has been going on for years too. DEC was one of the first to do it when some bright spark in their marketing department decided that their 64K memories were really 65K memories!
-
It's a bit like the old advertising con of pricing in "guineas (http://en.wikipedia.org/wiki/Guinea_%28British_coin%29#Replacement_by_the_pound)" rather than pounds.
-
Personally, I think everything should be rated in true Metric Gigabytes rather than Binary Gigabytes. (10, 100, 1000, etc).
But the capacity rating should be representative of the actual space available on the device.
However,
Technically the FAT table, for example, is real, and takes up real space on your device.
You can, however, also adjust the size of the FAT table to free up additional space if you wished.
The bigger the block sizes, the smaller the FAT table, but also the more wasted space on the device at the beginning and end of files.
If you only wished to put a single file on the device, you could essentially eliminate the FAT table.
I'm having troubles imagining what other overhead would be on the device.
Perhaps it would be like considering ECC memory as having 9 bit bytes rather than 8 bit bytes, so multiplying the capacity by 9/8, and selling 1 GB of ECC memory as 1.125GB.
-
Personally, I think everything should be rated in true Metric Gigabytes rather than Binary Gigabytes. (10, 100, 1000, etc).
Don't you mean decimal rather than metric? Personally, I think it would be much simpler is we all counted in hexadecimal [:D]
-
I think it would be much simpler is we all counted in hexadecimal
This guy prefers base 12 ...
-
I think it would be much simpler is we all counted in hexadecimal
This guy prefers base 12 ...
I'm not impressed. I've got eight on each hand.
-
One of the most famous computers ever built ENIAC worked in base 10.
-
One of the most famous computers ever built ENIAC worked in base 10.
Pity it didn't know about EDSAC.
Believe it or not, I used to keep some of EDSAC's mercury delay lines (the equivalent of to-days storage units) in a cupboard.
-
What happened to them?
I still have a nice bit of a serial computer, was the last unit I worked on of it's type. I was going to fix the fault, but had plenty of spare boards. It had a PSU fault that applied the unregulated bus to the 5V line. I replaced pass transistors, drivers, reference and overvoltage trip ( that did not work) and it worked again. It still ran and passed functional test even though Vcc was 14V, only it failed a single check. One faulty PM5403 was the cause of that, it popped it's lid.
-
What happened to them?
I've no idea. That was when I was the tech in the Dept. of Computing Science at Glasgow Uni. a very long time ago. They might still be there, but more than likely some spoilsport got rid of them because they were considered an environmental hazard.
-
Sean B
A PSU that outputs 14v onto the 5v line sounds rather a menace, did it do any damage ?
PS ENIACS designers knew all about base 2 but a base 10 design used fewer vacuum tubes.
-
PS ENIACS designers knew all about base 2 but a base 10 design used fewer vacuum tubes.
Interesting Syhprum. I better go and take a look at its architecture.
-
Looking at ENIAC, I suppose they would reduce the number of tubes, but only because they didn't have to use resources to convert between binary and decimal. If they had done all the computations in binary (which would have been a bit impractical) and assuming I understand the architecture properly, I believe they would have required even fewer tubes.
-
Memory sticks and hard drives for that matter are sold like broadband. We give you up to 20MB bandwidth actually means that 99.9% of customers are lucky to get close to 8MB. The only people that get the full bandwidth are those that live in a tent no more than 4 inches from the exchange and are fed by a platinum cable immersed in a bath of liquid nitrogen that’s wrapped in the finest yak wool hand stitch by Tibetan monks. Never met one myself but they must exist.
-
The 20M broadband is actually " We give you 20M from your house to the DSLAM, but there we throttle based on an alogrithm that uses time of day, how much your neighbour has used, our profit for the day/month/second and how much we are spending on our nice new office park, plus a factor based on how we feel". Would love to have the 8M broadband, but...............
@syhprum, it all survived the experience, but only one chip failed somewhat, with no ill effects. I replaced the power supply components and it worked well enough, ran well and passed the ATE afterwards. Was going to repair the board, but those were in stock in large numbers, as they rarely failed, so never did. I had done work on it before for another fault, capacitors that failed after 40 years of use.
-
Well I dug another USB pen out of the draw and it says it is FAT16 formatted rather than FAT32. Why is that? The properties on this one indicate that out of a capacity of 124.6Mb, 7.3Mb is used and 117.3Mb is available. That's much less than 10% unusable space.
-
FAT16 is limited to 2GB maximum.
FAT32 apparently has a minimum capacity of 512MB, but probably doesn't need to be used for capacities less than 2GB.
You can also format with NTFS, or various Linux formats.
Here are some notes on FAT16 vs FAT32.
http://technet.microsoft.com/en-us/library/cc940351.aspx
You should be able to calculate the FAT table size.
124.6MB / 8KB per entry = 15949 entries * 2 bytes / entry
Hmmm, that gives you about a 32KB FAT table.
8 GB / 4KB per entry * 4 bytes per entry = 8 MB
Oh, and redundant FAT tables, so multiply the above by 2.
Well, that certainly isn't enough to be the difference.
----------------------------------------------------------------------
NOW THAT IS ODD!!!
I found an 8GB USB keydrive. Looking at it with LINUX (sorry, no MSFT)
Dolphin reports it as 7.7 GB
Nautilus reports it as 8.2 GB
df reports it as 8040844 blocks. Divide that by 10242, and one gets about 7.66 GB.
fsck reports 2010211 clusters, multiply by 4 and we get 8040844 kb as above.
Ok,
I ran: fdisk -l
Disk /dev/sdb: 8253 MB, 8253341696 bytes
255 heads, 63 sectors/track, 1003 cylinders, total 16119808 sectors (I think it should be: 16113195, see below)
Units = sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0x86ee0675
Device Boot Start End Blocks Id System
/dev/sdb1 * 63 16113194 8056566 b W95 FAT32
Ok, I think that 8253 MB is WRONG [xx(]
Taking the byte size: 8253341696 bytes / 10243 = 7.68 GB
It looks like the first 64 blocks are wasted (FAT Table?)
[:-\]
Anyway, looking at the formatting.
By representing the USB keydrive as a hard drive.
they run into a few limits.
In particular, 255 heads and 63 sectors appears to be a major problem, and may be a throwback to DOS, although it isn't clear to me why they lost the ability to have 256 heads & 64 sectors. I wouldn't be surprised if that extra "head/sector" is physically in the memory, but not accessible. 1003 cylinders rather than 1024 cylinders appears to be cheating in capacity.
http://tldp.org/HOWTO/Large-Disk-HOWTO-4.html
The big question is whether that extra memory is actually on the keydrive, and not accessible. Perhaps I'll find a keydrive that I'm comfortable opening up. It is likely that they used an even number of "standard" chips to create the device. And there is probably some firmware on the device that does stuff like address translations.
-
The CHS numbers are a result of having just 3 bytes to store the number in, with heads being a value of 1 to 255, tracks being 1 to 63 and cylinders being a value from 1 to 1023. Any higher number needs an extra bit, which is not available. It is part of the legacy of computing, from the days when a PC shipped and ran with 64K of main memory. See http://en.wikipedia.org/wiki/Master_boot_record for more details.