a problem about memory
richi
Sorry, I am a new fish here and my english is not good!
I found a problem about memory. In a function I malloced more memory
than beforethat is 4.5625Kb in theory. Then I watched the memory size
of the function process in Linux and Unix through the command of "ps".
After comparing with the early memory size of the function which didn't
have change, I found the increaseded size was different! In Linux and Unix
the size is about 4Kb, but in windows the size watched from task manager
is far more than 4Kb.
can anyone help me and explain the phenomenon,please?
Thanks in advance!
Barry Margolin
wrote:
When a process starts up it gets a data segment allocated for it, and
malloc() uses this segment. Malloc() only needs to increase the size of
the data segment if you ask for more memory than is available. In your
case, there must already have been 4.5 KB available, so it didn't need
to grow the process.
--
Barry Margolin, bar...@alum.mit.edu
Arlington, MA
*** PLEASE post questions in newsgroups, not directly to me ***
*** PLEASE don't copy me on replies, I'll read them in the group ***
richi
"Barry Margolin" <bar...@alum.mit.edu> wrote in message
news:barmar-C36649....@news.motzarella.org...
> In article <gbfniv$tsi$1...@news.cn99.com>, "richi" <mhfc...@163.com>
> wrote:
>
>> Hi everyone,
>>
>> Sorry, I am a new fish here and my english is not good!
>>
>> I found a problem about memory. In a function I malloced more memory
>> than beforethat is 4.5625Kb in theory. Then I watched the memory size
>> of the function process in Linux and Unix through the command of "ps".
>> After comparing with the early memory size of the function which didn't
>> have change, I found the increaseded size was different! In Linux and
>> Unix
>> the size is about 4Kb, but in windows the size watched from task manager
>> is far more than 4Kb.
>>
>> can anyone help me and explain the phenomenon,please?
>>
>> Thanks in advance!
>
> When a process starts up it gets a data segment allocated for it, and
> malloc() uses this segment. Malloc() only needs to increase the size of
> the data segment if you ask for more memory than is available. In your
> case, there must already have been 4.5 KB available, so it didn't need
> to grow the process.
I had added several pointer variable in a function, the memory size that
the variable point to is 4.5625Kb in theory. Then I ran the programme in
Unix, Linux and Windows. Watched the process info, I found the the increased
memory size in Unix was the same in Linux. But it was different from the
size in Windows, the size in Windows was more than size in Unix and Linux.
The increased size in Unix and Linux was 4KB, and the increased size in
Windows was not constant. it was 8KB, 12KB, 16KB or 20KB.
Why this case happened?
I think that the increased memory size shoud be the same in the several
operating system. Could you explain the phenomenon for me?
Thank you very much!
Ian Collins
Because different platforms use different allocation schemes. Try a
windows group if you want to know more about that platform's behaviour.
--
Ian Collins.
Barry Margolin
wrote:
> I had added several pointer variable in a function, the memory size that
> the variable point to is 4.5625Kb in theory. Then I ran the programme in
4.5625 KB? Are you really measuring the size down to a half a byte?
Ian Collins
> In article <gbhfcj$cmf$1...@news.cn99.com>, "richi" <mhfc...@163.com>
> wrote:
>
>> I had added several pointer variable in a function, the memory size that
>> the variable point to is 4.5625Kb in theory. Then I ran the programme in
>
> 4.5625 KB? Are you really measuring the size down to a half a byte?
>
--
Ian Collins.
William Pursell
> > wrote:
>
> >> I had added several pointer variable in a function, the memory size that
> >> the variable point to is 4.5625Kb in theory. Then I ran the programme in
>
> > 4.5625 KB? Are you really measuring the size down to a half a byte?
>
> 4.5625 * 1024 = 4672 bytes.
I think the current nomenclature is:
4.5625Kb == 4562.5 bits
4.5625KB == 4562.5 bytes
4.5625KiB == 4672 bytes
Things would be so much more clear if we just
dropped the greek prefixes and switched to base 7. :)
Geoff Clare
> I think the current nomenclature is:
> 4.5625Kb == 4562.5 bits
> 4.5625KB == 4562.5 bytes
> 4.5625KiB == 4672 bytes
No, it's
4.5625kb == 4562.5 bits
4.5625kB == 4562.5 bytes
4.5625KiB == 4672 bytes
AFAIK there is no formal standard for the uppercase K prefix.
Historically K was chosen to mean 1024 in order to distinguish
it from k which meant 1000. So I think most people would still
take K to mean 1024.
--
Geoff Clare <net...@gclare.org.uk>
Robert Harris
represented multipliers of 10 to a positive integer power, whereas lower
case letters (mVolt, etc.) represented multipliers of 10 to a negative
integer power. So, for example, M is mega and m is milli. In this schma,
kilo is a capital K.
This is surely standardised somewhere!
Robert
Rainer Weikusat
The usual abbreviation for kilometer is km and for kilogram, it is
kg.
Ralf Fassel
| When I was at school, I was taught that capitals (Kbits, MB, etc.)
| represented multipliers of 10 to a positive integer power, whereas
| lower case letters (mVolt, etc.) represented multipliers of 10 to a
| negative integer power. So, for example, M is mega and m is
| milli. In this schma, kilo is a capital K.
'kilo' always uses a lower case 'k'...
http://physics.nist.gov/cuu/Units/prefixes.html
<quote>
Because the SI prefixes strictly represent powers of 10, they
should not be used to represent powers of 2. Thus, one kilobit, or
1 kbit, is 1000 bit and not 210 bit = 1024 bit. To alleviate this
ambiguity, prefixes for binary multiples have been adopted by the
International Electrotechnical Commission (IEC) for use in
information technology.
</quote>
=> http://physics.nist.gov/cuu/Units/binary.html
Mainstream usage is different, though (image your favourite computer
store offering 2048MebiByte-RAM laptops...:o)
R'
Eric Sosman
> Robert Harris <robert....@blueyonder.co.uk> writes:
>> [...]
>> When I was at school, I was taught that capitals (Kbits, MB, etc.)
>> represented multipliers of 10 to a positive integer power, whereas lower
>> case letters (mVolt, etc.) represented multipliers of 10 to a negative
>> integer power. So, for example, M is mega and m is milli. In this schma,
>> kilo is a capital K.
R.H. was mis-taught. The upper- and lower-case mnemonic
*almost* holds: All upper-case prefixes are magnifiers, and
all "minifiers" are lower-case (yes, µ is a lower-case letter).
However, the abbreviations k, h, and da (kilo, hecto, deka; 10^3,
10^2, 10^1) are all magnifiers despite their case.
> The usual abbreviation for kilometer is km and for kilogram, it is
> kg.
A purely technical quibble: The kilogram is a fundamental
SI unit, so "1 kg" means "one kilogram" rather than "one thousand
grams." When using larger or smaller units, though, one never
combines another prefix with the symbol kg; one behaves as if
stating quantities in the non-fundamental unit gram, g. Thus,
a tonne is 1 Mg, not 1 kkg.
http://physics.nist.gov/cuu/Units/prefixes.html
When I was a kid we were taught two systems of metric units:
MKS and CGS, for meter-kilogram-second and centimeter-gram-second,
naming the principal base units of the respective systems. MKS
was said to be favored by physicists and astronomers, CGS was the
choice of chemists and biologists (Wikipedia says otherwise, but
that's what I was told at the time). The modern SI seems to be a
formalization of MKS; perhaps the chemists' acid-tongued oratory
offended the other voting blocs. (It always perplexed me that both
MKS and CGS involved base units that appeared already multiplied; I
kept wondering why the obvious MGS system was never adopted.)
Still, CGS lives on in an old geeky cartoon. I won't attempt
to reproduce it in ASCII art, so here's a description: At the top
of the picture is the word "erg," the CGS unit of energy. In the
middle there's a drawing of a bug lying on its back with its very
many feet waving helplessly in the air. And below is the explanation:
"An erg is a dyne centimeter."
Rainer Weikusat
'Mainstream usage', ie 'usage by people not accustomed to deal with
binary computers' is to use powers of ten as 'usual orders of
magnitude'. That's just really silly when using binary computers,
because calculating powers-of-ten requires fairly complicated
algorithms, while calculating powers-of-two can be done with simple
shift operations.
Barry Margolin
Eric Sosman <Eric....@sun.com> wrote:
> Rainer Weikusat wrote:
> > Robert Harris <robert....@blueyonder.co.uk> writes:
> >> [...]
> >> When I was at school, I was taught that capitals (Kbits, MB, etc.)
> >> represented multipliers of 10 to a positive integer power, whereas lower
> >> case letters (mVolt, etc.) represented multipliers of 10 to a negative
> >> integer power. So, for example, M is mega and m is milli. In this schma,
> >> kilo is a capital K.
>
> R.H. was mis-taught. The upper- and lower-case mnemonic
> *almost* holds: All upper-case prefixes are magnifiers, and
> all "minifiers" are lower-case (yes, µ is a lower-case letter).
> However, the abbreviations k, h, and da (kilo, hecto, deka; 10^3,
> 10^2, 10^1) are all magnifiers despite their case.
The upper/lowercase distinction is used for the initials that would be
ambiguous because there are two prefixes that start with the same
letter, e.g. milli and mega. For the ones that aren't ambiguous, either
case is acceptable.
Eric Sosman
> In article <1222788735.185228@news1nwk>,
> Eric Sosman <Eric....@sun.com> wrote:
>
>> Rainer Weikusat wrote:
>>> Robert Harris <robert....@blueyonder.co.uk> writes:
>>>> [...]
>>>> When I was at school, I was taught that capitals (Kbits, MB, etc.)
>>>> represented multipliers of 10 to a positive integer power, whereas lower
>>>> case letters (mVolt, etc.) represented multipliers of 10 to a negative
>>>> integer power. So, for example, M is mega and m is milli. In this schma,
>>>> kilo is a capital K.
>> R.H. was mis-taught. The upper- and lower-case mnemonic
>> *almost* holds: All upper-case prefixes are magnifiers, and
>> all "minifiers" are lower-case (yes, µ is a lower-case letter).
>> However, the abbreviations k, h, and da (kilo, hecto, deka; 10^3,
>> 10^2, 10^1) are all magnifiers despite their case.
>
> The upper/lowercase distinction is used for the initials that would be
> ambiguous because there are two prefixes that start with the same
> letter, e.g. milli and mega. For the ones that aren't ambiguous, either
> case is acceptable.
Not according to SI. Tera is T, not t, giga is G, not g,
femto is f, not F. Also, deci and deka are d and da, not d and D.
Of course, "acceptable" is defined by the acceptor or
rejector: You're at liberty to "accept" whatever you like. As
long as you don't insist that everyone else accept your abomin--
er, idiosyncrasies, it matters not a pp[*] to me.
[*] picopenny, an amount of money whose value has dwindled
somewhat from the effects of inflation.
--
Eric Sosman
eso...@ieee-dot-org.invalid
Barry Margolin
Eric Sosman <eso...@ieee-dot-org.invalid> wrote:
> Barry Margolin wrote:
> > In article <1222788735.185228@news1nwk>,
> > Eric Sosman <Eric....@sun.com> wrote:
> >
> >> Rainer Weikusat wrote:
> >>> Robert Harris <robert....@blueyonder.co.uk> writes:
> >>>> [...]
> >>>> When I was at school, I was taught that capitals (Kbits, MB, etc.)
> >>>> represented multipliers of 10 to a positive integer power, whereas lower
> >>>> case letters (mVolt, etc.) represented multipliers of 10 to a negative
> >>>> integer power. So, for example, M is mega and m is milli. In this schma,
> >>>> kilo is a capital K.
> >> R.H. was mis-taught. The upper- and lower-case mnemonic
> >> *almost* holds: All upper-case prefixes are magnifiers, and
> >> all "minifiers" are lower-case (yes, � is a lower-case letter).
> >> However, the abbreviations k, h, and da (kilo, hecto, deka; 10^3,
> >> 10^2, 10^1) are all magnifiers despite their case.
> >
> > The upper/lowercase distinction is used for the initials that would be
> > ambiguous because there are two prefixes that start with the same
> > letter, e.g. milli and mega. For the ones that aren't ambiguous, either
> > case is acceptable.
>
> Not according to SI. Tera is T, not t, giga is G, not g,
> femto is f, not F. Also, deci and deka are d and da, not d and D.
So if I wrote that I have 10 tb disk space, you would have trouble
understanding it? Since there is no lowercase t prefix, what else could
it mean?
> Of course, "acceptable" is defined by the acceptor or
> rejector: You're at liberty to "accept" whatever you like. As
> long as you don't insist that everyone else accept your abomin--
> er, idiosyncrasies, it matters not a pp[*] to me.
According to the dictionary, the only correct spelling of problem is
"problem". Although spelling it "problm" won't win you a spelling bee,
you'll still be understood. That's what I mean by "acceptable".
William Pursell
> In article <pOOdnUYiQJai5H7VnZ2dnUVZ_uadn...@comcast.com>,
> Eric Sosman <esos...@ieee-dot-org.invalid> wrote:
>
>
>
> > Barry Margolin wrote:
> > > In article <1222788735.185228@news1nwk>,
> > >>>> [...]
> > >>>> When I was at school, I was taught that capitals (Kbits, MB, etc.)
> > >>>> represented multipliers of 10 to a positive integer power, whereas lower
> > >>>> case letters (mVolt, etc.) represented multipliers of 10 to a negative
> > >>>> integer power. So, for example, M is mega and m is milli. In this schma,
> > >>>> kilo is a capital K.
> > >> R.H. was mis-taught. The upper- and lower-case mnemonic
> > >> *almost* holds: All upper-case prefixes are magnifiers, and
> > >> However, the abbreviations k, h, and da (kilo, hecto, deka; 10^3,
> > >> 10^2, 10^1) are all magnifiers despite their case.
>
> > > The upper/lowercase distinction is used for the initials that would be
> > > ambiguous because there are two prefixes that start with the same
> > > letter, e.g. milli and mega. For the ones that aren't ambiguous, either
> > > case is acceptable.
>
> > Not according to SI. Tera is T, not t, giga is G, not g,
> > femto is f, not F. Also, deci and deka are d and da, not d and D.
>
> So if I wrote that I have 10 tb disk space, you would have trouble
> understanding it? Since there is no lowercase t prefix, what else could
> it mean?
Harware with tuberculosis is bad...but 10 cases simultaneously
is probably fatal. Actually, I would assume you meant
10 * 2^40 bytes, but be uncertain that you might have only
meant 10 * 2^37 bytes (10 * 2^40 bits). But if you were a marketing
person, I would assume you meant 10 ^ 13 bits (since they
prefer bits because they can inflate the number). The
whole system is hopelessly ambiguous at this point, since
you never know if an engineer who understands the binary
language of moisture vaporators wrote the spec, or
a marketing wonk. And then there are plenty of allegedly
technical people who don't understand binary at all.
Ralf Fassel
| And then there are plenty of allegedly technical people who don't
| understand binary at all.
There are 10 kinds of people: those who understand binary,
and those who don't.
R'
Eric Sosman
> In article <pOOdnUYiQJai5H7V...@comcast.com>,
> Eric Sosman <eso...@ieee-dot-org.invalid> wrote:
>
>> Barry Margolin wrote:
>>> The upper/lowercase distinction is used for the initials that would be
>>> ambiguous because there are two prefixes that start with the same
>>> letter, e.g. milli and mega. For the ones that aren't ambiguous, either
>>> case is acceptable.
>> Not according to SI. Tera is T, not t, giga is G, not g,
>> femto is f, not F. Also, deci and deka are d and da, not d and D.
>
> So if I wrote that I have 10 tb disk space, you would have trouble
> understanding it? Since there is no lowercase t prefix, what else could
> it mean?
so if i rot that i hv tentb disk u wd hav trubbl u11g it?
since thers 0 lc t prefix wht else cd it mean?
In other words, standards for nomenclature and terminology
exist for the purpose of reducing the chance for misunderstanding,
for increasing the chance that a message carries the same meaning
when received as it had when uttered. In your example, I would
most likely guess that "t" probably meant "T," but a person who
requires his readers to guess at his meaning should consider
dropping engineering and taking up poetry.
Plus, I'd still be uncertain whether you meant "T" or "Ti"
(do you have 10 Tb or 10 Tib ~= 11 Tb?). In the context of a disk
I'd guess that "b" most likely meant "B," and be thankful you
weren't writing about a networking component where "b" might
mean "bit" instead.
>> Of course, "acceptable" is defined by the acceptor or
>> rejector: You're at liberty to "accept" whatever you like. As
>> long as you don't insist that everyone else accept your abomin--
>> er, idiosyncrasies, it matters not a pp[*] to me.
>
> According to the dictionary, the only correct spelling of problem is
> "problem". Although spelling it "problm" won't win you a spelling bee,
> you'll still be understood. That's what I mean by "acceptable".
It's good that messages can resist some noise. It's bad to
inject gratuitous noise.
--
Eric Sosman
eso...@ieee-dot-org.invalid