First of all let's see the final result. Probably the laws of your country already prescribed the use of the following system, even if you haven't heard of it.

No exception, no excuse, this is the only official system. The prefixes are derived from greek words, following the scientifical manner, so they have to be pronounced by the generic latin rules. Not guy-buy, say ghee-bee. Notice that the correct abbreviation of kilo is a lower case k, but of kibi is beginning with an upper case K.

The decimal prefixes are based on 10 raised to integer power, 1000 equals to 10³. The binary prefixes are based on 2, 1024 equals to 2¹⁰. (3 GiB = 3×2³⁰ B, 3×1024³ B, 3×2²⁰ KiB, 3.221 GB)  :-)  This is why you better use this little tool to make calculations like these.

Note: you mustn't combine prefixes, no kilomega and others are allowed.

In the 50s the size of operative memory in the computers of that time became bigger and bigger, yet even some thousand bytes. This enlargement required abbreviations, so kilo, later mega offered themself. But the memory sizes, because of technical factors, are always a product of 2s. This is why computer engineers choosed 2¹⁰ for a larger unit for binary arrays, this is 1024, it's very close to 1000, so kilobytes or kilobits became standing for 1024 bytes or bits. This alternative use of the prefixes in special cases were well known, as good as a standard. In 1986 the ANSI and the IEEE issued this method as a real standard, and everybody was happy.

Almost everybody. When you bought a 3.5 inch size floppy disk, you saw: "1.44 MB". Well, the size was 2×80×18×512 bytes, which is 1.44×1024×1000 bytes. Interesting, isn't it? And the chaos had rescrudesced. A CD is 700 MB big, 700×1024², 734 million bytes. A DVD is 4.7 GB big, 4.7×1000³, 4700 million bytes. How many bytes did you get in a 320 GB size hard disk? Yes, 298 thousand million bytes. On the level of the tera sizes the deviation is exactly 10 percents. The only certain thing was the size of the electronic memories (RAM) because those always have to be measured in binary exponent quantities.

The Institute of Electrical and Electronic Engineers (IEEE), International Organization for Standardization (ISO) and International Electrotechnical Commission (IEC) created and proposed the new system of binary prefixes, look at the first table again. Following some other directions now the ISO/IEC 80000-13:2008 international standard lays down the rules clearly. This procedure induced reissuing regional or national standards, too. The Bureau International des Poids et Mesures, the office of the SI measuring unit system now definitely prohibits the applying of the original decimal prefixes at any other situation than with the standardized SI units.

So, maybe you have never heard about it, maybe you thought it only an alternative practice, but if your country has already enactmented this system, then you are bound to use this. No exception, no excuse. At least you would be bound to.

But when? Always, if you want to mean powers of 2 with prefixes. And never if you mean powers of 10. 1 kB is 1000 bytes, 1 KiB is 1024 bytes, that's all.

And other people? Ehm... You have to guess at what they thought. Here 1 kilobyte means 1024 bytes, there 1000. Only a few softwares use the new official standard, this makes hard to estimate real file sizes, free spaces. Data transfer capacity is usually measured truly on decimal base, 100 kbps really means 100,000 bits per second.

We can't tell you where you have to calculate with 1 MB as 1 megabyte and where as 1 mebibyte. But if you have to calculate that, it's easier to do with this little program.