May you provide any reference that authors of time or ulimit use SI units? The Wikipedia page I linked the 1024 byte unit is still widely used as "kilobyte".

You are misreading the information provided to you. Using kilobyte for a unit of 1024 bytes is WRONG, no matter how many hardware vendors are still using it wrong - that doesn't make it any more correct.

Folks who discover software or manpages using the wrong term should file a bug report so it can be fixed.

@ArturMeinild I do not care about right and wrong. What I do care about is that there are people which use "kilobyte" as 1024 bytes and they may be authors of ulimit and time.

@abukaj fount the source for GNU time :)

While this answer is correct, the realty is that it's extremely common to use the word kilobyte to mean 1024. The 'kibi' prefix ('gibi', 'mibi') is only just beginning to become common and I'm not sure it will ever fully displace the 'wrong' terminology partly because it sounds silly and there's no practical need to think in terms of true kilobytes. I understand why it's wrong but this pedantry has created more confusion than value, IMO. Maybe it helps a student in their first week in CS or on the job but beyond that, it's pretty irrelevant.

"No, kilobytes is 1000. We follow the International System of Units so the prefix "kilo" refers to 1000 not 1024. Any other reference is wrong." Who is "we"? Kilo was the only prefix until 2008ish (when IEC 80000-13 defined kibi-), and it almost always meant 1024 in a computing context, so it's kind of presumptuous to say existing utilities, using the accepted definition of "kilobyte" in use from the first time we needed to describe a kilobyte of memory, through 2008, are retroactively wrong.

I think it's fine to keep that detail on KiB vs. kB, but you should actually answer the question first, rather than insisting on pedantic correctness first and burying the actual answer.

"Kilo was the only prefix until 2008ish [...], and it almost always meant 1024 in a computing context, ..." - No, it didn't. It meant 1000 in hard drive capacity, clock speeds (a 3GHz CPU is 3 000 000 000 Hz), and transfer speed (Fast Ethernet is 100 000 000 bits/sec). It meant 1024 in the context of RAM, flash, and low level programming / system software (filesystem blocks, memory pages, etc). Most everywhere else, it was a coin flip. The situation has been a mess for a long time.

@ArturMeinild It's not wrong. What's happened is that a standard was established which conflicts with common usage in software - usage which vastly predates the standard, since it was first seen in print in 1959. It doesn't follow the more modern ISO standard, sure. But in the field of software and electronics when counting power-of-2 quantities, it's a perfectly valid and correct meaning, because they largely still do not follow that ISO standard.

@ShadowRanger what now? my answer is in order of editing. Nothing wrong with that. And to all: comments are not for discussions.

@JimmyJames perhaps in casual discussion, but programmers are expected to know the difference.

@user253751 I expect a developer to understand kilobyte to be 1024 in relevant contexts.

@JimmyJames I expect them to know that kilobyte is ambiguous, has always been, and then (if they have the ability) change it to kibibyte which is unambiguous.

@marcelm: A "1.44MB" floppy holds 1.44 times 1,024,000 bytes, but even now there's no name for that unit of storage even though I think that for many purposes involving large numbers of fungible chunks of power-of-two sizes of ~4,096 bytes or less, the most useful terminology would be power-of-1000 multiples of 1024, so something twice as big as a 720kiB thing would be a 1.44_B thing.

40+ year old here, and a megabyte will be 1024 bytes until I die.

@THEJOATMON you might want to keep your megas and kilos straight or you'll be even more messed up than you realize. I consider the power of 2 definition to be short sighted no matter how convenient it was when it came about. As the quantities get larger, the differences get greater; a KiB is 2% larger than a KB, but a TiB is 10% larger than a TB.

@Graham the definitions of the K, M, etc. prefixes were well established before any standards bodies adopted them. They go back to the establishment of the metric system.

Who is this "we" that uses SI units? I don't know of anywhere in the world that uses entirely SI units. At least, I've never seen a speed limit sign in kilometers per kilosecond.

The moves to reduce this ambiguity gained pace in 1999 with IEC advocating the Ki-etc terminology, and again in 2009 with ISO 80000. SI is a red-herring, it isn't an SI core unit nor in the very long list of acceptable units to use "alongside" in any of their brochures. It is not their concern because it is outside the scope they defined for themselves.