Re: how to test this piece of C code

Randy Howard said:

On Sat, 10 Mar 2007 03:21:20 -0600, Richard Heathfield wrote
(in article <sOCdnRu3b6LV5W_YRVnyuAA@xxxxxx>):

subramanian100in@xxxxxxxxx, India said:

On Mar 10, 3:27 am, Richard Heathfield <r...@xxxxxxxxxxxxxxx> wrote:
subramanian10...@xxxxxxxxx, India said:

I do not understand this. Kindly explain. Does it mean that in real
world applications SIZE_MAX memory won't be asked from
malloc/realloc

No, it doesn't mean that.

or there won't be 4GB of input file for testing ?

No, it doesn't mean that, either. It just means that most companies
are unlikely to have a machine containing that much physical RAM.

True enough, unless you happen to be working for a platform vendor or
RAM vendor. I've had access to max-mem configs many times over the
years, most of them while working for hardware manufacturers, not pure
software shops.

Furthermore, it will become /less/ true as the years roll by.

Nowadays, it is still unusual (but not exactly rare) for machines to
have 1GB of RAM. Give it a few years, and it will become unusual (but
not exactly rare) to find a machine with /less/ than 1GB of RAM. And
4GB is only two binary orders of magnitude away.

For comparison purposes: In 1990 (when I bit the bullet and actually
bought one of these suckers), my machine had 4MB of RAM (which was
tons!), and 80MB of hard disk space (which, I calculated, would last me
20 years - cough). Seventeen years on, my principal machine has 256MB
of RAM (up by six binary orders of magnitude) and 25GB of hard disk
space (up by over eleven binary orders of magnitude), and is probably
considered low-spec by many here.

Past results are no guide to future performance, they say - but if they
were, then in 2024 I could expect to be enjoying the benefits of a
machine with 16GB of RAM and over 50TB of disk space - and it will
/still/ be considered low-spec!

--
Richard Heathfield
"Usenet is a strange place" - dmr 29/7/1999
http://www.cpax.org.uk
email: rjh at the above domain, - www.
.

Relevant Pages

  • Re: how to test this piece of C code
    ... RAM vendor. ... probably be, relatively speaking, about the same as a 64GB config is ... 4GB is only two binary orders of magnitude away. ...
    (comp.programming)
  • Re: factorial of very big number like 4096
    ... There are libraries that do arithmetic on quantities whose magnitude ... it's good enough for many applications that don't *really* need ... Supposing 8GB main RAM, ... bits or 2 796 202 decimal digits. ...
    (comp.lang.c)
  • Re: OT - Stupid Woodworking Mistake
    ... mea culpa. ... (still having a hard time thinking of RAM in the magnitude of ... when I upgades the RAM to 64K it died horribly. ...
    (rec.woodworking)
  • Re: how fast is A1
    ... something like 8 KBytes of RAM. ... magnitude better than AOS4! ... I could use the switches on my IMSAI 8800 to enter in an OS in less than 1K ...
    (comp.sys.amiga.advocacy)