Standards body tries to change the vernacular

Electronics
#1

Continuing the discussion from You don't need an Arduino to Flash an LED:

In looking at the link from @zmetzing post I came across some prefixes I hadn’t heard before; kibi, mebi, and gibi.

These are defined in ways I have always seen kilobit, megabyte, and gigabyte. Granted these latter terms seem to have had varying definitions. For instance, I have seen megabyte refer to 1,000,000 bytes as well as 1,048,576 bytes.

So I am curious, has any else seen and more importantly used these new prefixes?

http://physics.nist.gov/cuu/Units/binary.html

#2

Yes, but only sort of.
Textbooks occasionally use the “new math”, but honestly, it’s largely ignored in the classes I’ve been exposed, other than to explain the “new prefixes”.

#3

Being a recovering physicist, I’d come across these before. Physicists at both Fermilab and CERN seemed to concern themselves with the differences due to the volume of data they generated, transferred, and stored. Both are amazing facilities, by the way.

Now that I’m an engineer: eff it, close enough!

This is me and my fellow physics folks from SMU inside of ATLAS at CERN. ATLAS is one of the two general purpose detectors on the LHC at CERN.

8 Likes
#4

Absolutely. They’re the base roots for RAM, storage, and throughput in cloud computing.

Their increase in common usage stems from a number of lawsuits about the topic:

If you were to visit our friends at Amazon Web Services, you’ll see that they quote their RAM in gibibytes, not the more traditional gigabytes that people are used to seeing:

https://aws.amazon.com/ec2/instance-types/

Interestingly, this has resulted in the expansion of metrics that you have to understand in order to correctly size an instance:

Gibibytes for RAM
Gigabytes for Storage
Gigabits per second for throughput

#5

This reference may be useful:

7 Likes
#7

I’ve seen them used in some open source operating system tools.

I vaguely recall seeing them in a datasheet (maybe a communications chip).

I’ve started using them for online posting.

#8

But 1.00 MΩ is not 1048576 ohms.

1 Like
#9

Could be if it’s a 5% tolerance resistor.:wink:

3 Likes
#10

I’ve come across this many times over the years going pretty far back (as long as I can remember using Linux, a decade plus). Usually related to data storage utilities and data storage software. For example “df -h” and “df -H” do the same but one uses base 1000 and one uses base 1024. From the man.

       -h, --human-readable
              print sizes in powers of 1024 (e.g., 1023M)

       -H, --si
              print sizes in powers of 1000 (e.g., 1.1G)

A lot of data storage software specifically uses IEC units (base 1024) like TiB to avoid any confusion. For example “fdisk -l” output.

Disk /dev/sdj: 3.7 TiB, 4000787030016 bytes, 7814037168 sectors
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 4096 bytes
I/O size (minimum/optimal): 4096 bytes / 4096 bytes
Disklabel type: gpt

I’m glad there’s this distinction so you can have meaningful conversations with people about storage and not get wrecked by the storage device manufacturer size shenanigans (SI instead of IEC). If we could undo everything and make storage device manufacturers advertise the actual size in base 1024 that would be great, but at this point we make do with it.

6 Likes
#12

This is an interesting read.

3 Likes
#13

Hmmm. My 2-year old nephew uses these phrases all the time. Kids. So tech-savvy …

Seriously - thanks @wandrson and @bknapp. I learned something. Always a good thing.

4 Likes