Linux Mint Forums

Why are the size of hard drives larger than they are?

I bought a 4 terabyte which was really only 3.64 TiB!

Already when it had connected the o started my formatting program
(KDE partition manager )so "this" said that it was 3.64
and that was BEFORE I even chose the file system or partition table type.

If I can't access more than 3.64 regardless of which file system I choose -
why does the manufacturer indicate that it is 4 TiB? That's a lie !

Vilsen wrote: ⤴Thu May 16, 2019 5:56 am why does the manufacturer indicate that it is 4 TiB? That's a lie !

I'm sure they indicate 4 TB, not TiB. There's usually a disclaimer on the packaging defining that they are using a decimal based definition for that unit, as opposed to the binary basis for TiB. It's a marketing trick basically.

... and ....

The physical device is sized to hold the claimed 4TB. But, when formatted, the filesystem is suddenly a bit smaller than that claimed by the disk manufacturer. There are a number of reasons for this, but it all boils down to filesystem overhead and, as alluded to by gm10, the difference between a tera-byte (1024 * 1024 * 1024 bytes) and what the disk manufacturer is selling (1,000 x 1,000,000 bytes).

YUP - it's the Technical Dept Vs the Sales Dept


any drive is sold as it's Un-formatted size - - but you buy it Formatted.
- - so you do indeed, get an Smaller Drive.

and it's always been like that.


so your - 1Tb drive is actually 930Gb in size.
- 2Tb drive is actually 1860Gb in size.
etc etc ..
and my New 500Gb is actually 465Gb in size.

This sort of thing is routine, unfortunately.

Obviously manufacturers have an interest in having nominal designations sound as big as possible but, really, anyone who knows about these things knows what to expect and what those numbers mean.

If you go to the hardware store and told them they are lying because 2" x 4" lumber is not 2" x 4" they would laugh at you and rightly so.

I'm glad your answers were as I feared.
We live in a world where excesses persist.
But a question that I have not received an answer:

Why are all file system formatting with exactly the same result?

I have tried with NTFS and exFAT o others but all are 3.64 TB.

Does Apple / MacOS have any better or is it physically
IMPOSSIBLE to get more than 3.64 from a "4TB" disk?
Observe quotes - something they might start using ...

Looks like you didn't really understand any of our answers, unfortunately. The point is:

4 TB = 3.64 TiB

... and other than the TB vs. TiB issue, the thing that annoys me most about these even after 15 years or so still never ending threads is the "it's just marketing" answers.

No it is not.

The kilo-, mega-, giga-, tera- and so on prefixes are defined prefixes meaning times 1000, 1000^2, 100^3, 1000^4 and so on respectively. Not the binary 1024, 1024^2, 1024^3, 1024^4 and so on. Back in the days of kilo- no one cared due to the difference being tiny anyway but when sizes grew the difference between the two uses also grew. At a time when the binary prefixes didn't even exist as separate words drive manufacturers then did the only reasonable thing and adopted the standard, globally defined and only scientifically correct SI prefixes. Yes, I'm sure they didn't at the time mind that it makes drives seem larger in the eyes of consumers --- but that doesn't in fact matter. They are still right, and consumers still wrong.

Now there's of course wrong and wrong, and I can at times actually be persuaded to find the practical one more right but, quick, without looking it up, can your Gigabit network card shuffle 1000^3 or 1024^3 bits per second out the door? (spoiler, the example underwrites my next sentence). In fact try and come up with any hardware category other than memory technology that use the 1024 standard for prefixes. Having one do makes for a definite mess.

rene wrote: ⤴Fri May 17, 2019 6:21 am ... and other than the TB vs. TiB issue, the thing that annoys me most about these even after 15 years or so still never ending threads is the "it's just marketing" answers.

No it is not.

Of course it's marketing. Computers store data in binary patterns, so storage units have originally always been binary. Granted, using SI prefixes for the binary units wasn't ideal but the binary prefixes (like TiB) would not be standardized until later. Nevertheless, Windows is using binary units to this day, RAM manufacturers are using binary units to this day, it's only storage manufacturers that suddenly came up with the idea to use decimal units instead when originally they had not.

That last part is important, and it becomes even more clear that it was just a marketing ploy when you look at the exact history. As some of us will remember, there was such a thing as floppy disks. Now popular formats were 360 KB and then 720 KB double-density floppies, and here the units were still binary. Then came the 1.44 MB high density variant of the latter, and here the unit was suddenly a mix of binary and decimal, for they used MB = 1 024 000 bytes, or 1000 KiB (and here I believe the 1.2 MB 5 1/4 floppy was still fully binary but my memory is a bit hazy on that). Then eventually they decided to go fully decimal on hard drives and after a few court cases and resulting settlements we ended up with the disclaimer on them explaining the unit.

I still haven't got over the fact that the Chrysler 2.0 liter engine only had 1996 cc. Talk about misrepresenting your product! I bought the car in 1998 and I still haven't got over it. I was robbed! Robbed I tell ya!

GS3 wrote: ⤴Fri May 17, 2019 10:41 am I still haven't got over the fact that the Chrysler 2.0 liter engine only had 1996 cc. Talk about misrepresenting your product! I bought the car in 1998 and I still haven't got over it. I was robbed! Robbed I tell ya!

Look on the bright side old chap. After 20 years the bores on that engine will have worn so much the capacity has probably increased to 2016 cc. That's an extra 20 cc for free!

gm10 wrote: ⤴Fri May 17, 2019 8:43 am Now popular formats were 360 KB and then 720 KB double-density floppies, and here the units were still binary. Then came the 1.44 MB high density variant of the latter, and here the unit was suddenly a mix of binary and decimal, for the actual capacity was 1 024 000 bytes, or 1000 KiB (and here I believe the 1.2 MB 5 1/4 floppy was still fully binary but my memory is a bit hazy on that).

I wonder if this will be useful? From images made from floppies using dd, as shown in properties via Dolphin:
5¼" 360.0 KiB (368,640) - from the first disk of the Amstrad PC1640 System Disk set
5¼" 1.2 MiB (1,228,800) - from a freshly formatted disk

3½" 720.0 KiB (737,280) - at least I assume it's from one, will have to check the image, might be 5¼"
3½" 788.0 KiB (806,912) - from a disk with ElfBowl on it; I do think this one's a 3½" image
3½" 1.4 MiB (1,474,560) - from a freshly formatted disk

I don't know if I still have the images made from the various old formats which I created using the CatWeasel PCI card; I'll have to get a system up and running which this card will work on. These included all sorts including GCR etc. but had a hard disk failure years ago. One of these days I'll replace those caps on my AMD Athlon & have another go. Or find out why the old Dell system won't boot with its new compliment of 512MB RAM ..

GS3 wrote: ⤴Fri May 17, 2019 10:41 am I still haven't got over the fact that the Chrysler 2.0 liter engine only had 1996 cc.

This is (or was) done for tax class reasons, I believe. The first car I owned in this country was 988CC & the second 998CC.

cliffcoggin wrote: ⤴Fri May 17, 2019 5:45 pm Look on the bright side old chap. After 20 years the bores on that engine will have worn so much the capacity has probably increased to 2016 cc. That's an extra 20 cc for free!

Mine must have been about 2L by the time I'd done with them ..

GS3 wrote: ⤴Fri May 17, 2019 10:41 am I still haven't got over the fact that the Chrysler 2.0 liter engine only had 1996 cc. Talk about misrepresenting your product! I bought the car in 1998 and I still haven't got over it. I was robbed! Robbed I tell ya!

Interesting point. All 2.0 liter engines I have personally encountered actually have less than 2.0 liter displacement.

BG405 wrote: ⤴Fri May 17, 2019 7:18 pm

gm10 wrote: ⤴Fri May 17, 2019 8:43 am Now popular formats were 360 KB and then 720 KB double-density floppies, and here the units were still binary. Then came the 1.44 MB high density variant of the latter, and here the unit was suddenly a mix of binary and decimal, for the actual capacity was 1 024 000 bytes, or 1000 KiB (and here I believe the 1.2 MB 5 1/4 floppy was still fully binary but my memory is a bit hazy on that).

I wonder if this will be useful? From images made from floppies using dd, as shown in properties via Dolphin:
5¼" 360.0 KiB (368,640) - from the first disk of the Amstrad PC1640 System Disk set
5¼" 1.2 MiB (1,228,800) - from a freshly formatted disk

3½" 720.0 KiB (737,280) - at least I assume it's from one, will have to check the image, might be 5¼"
3½" 788.0 KiB (806,912) - from a disk with ElfBowl on it; I do think this one's a 3½" image
3½" 1.4 MiB (1,474,560) - from a freshly formatted disk

Haha, excellent, I was going by memory but I love me some hard data. Note that you are using IEC binary unit symbols for everything here but as the math shows, and confirming what I said above, for the MB notation they actually started using a mix of decimal and binary, for they used 1 MB = 1 000 KiB = 1 024 000 bytes whereas 1 MiB is actually 1024 KiB or 1 048 576 bytes yet the SI definition of 1 MB is 1 000 KB or 1 000 000 bytes. So this was the transitional phase where they were starting to blow up the numbers somewhat but hadn't fully committed.

Here's the math:

5¼" 360.0 KiB (368,640) = 360 * 1024
5¼" 1.2 MiB (1,228,800) = 1.2 * 1000 * 1024

3½" 720.0 KiB (737,280) = 720 * 1024
3½" 788.0 KiB (806,912) = 788 * 1024
3½" 1.4 MiB (1,474,560) = 1.44 * 1000 * 1024

You wrote:

" Looks like you didn't really understand any of our answers,
unfortunately. The point is: 4 TB = 3.64 TiB "


Well, I guess I wrote TiB somewhere but I expressed myself
foolishly.

Somewhere I read:



A tebibyte (TiB) is a unit of measure used to describe computing capacity. The prefix tebi comes from the binary system for measuring data capacity. That system is based on powers of two. One tebibyte equals 240 or 1,099,511,627,776 bytes.

Maybe if the HD-manufacturers insdtead used TEBIBYTES instead we did not need to
be CONFUSED ?????

Vilsen wrote: ⤴Sat May 18, 2019 8:10 am Maybe if the HD-manufacturers insdtead used TEBIBYTES instead we did not need to
be CONFUSED ?????

Hence the side-discussion in this thread. It's impossible for them to go back to using the binary units now unless the entire industry gets regulated at once though. Consumers only look at the bigger number, that's why that industry started using decimal units in the first place, so if one manufacturer alone started using binary units again they'd be at a serious disadvantage.