efficient code to do left shit on 64-bit values

[aurgathor]

I use BC 5.02 (long is 32 bit) and I wonder if
there's any efficient code that would allow me
to left shift 64 bit values that I currently use
this way:

typedef struct {
unsigned long lsLong;
unsigned long msLong;
} b64_struct;

This struct can be changed if needed.

The least significant part can contain any value
in the 0..2^32 -1 range. The most significant
part; however, will usually be 0, but it may
cintain some smaller values, usually under 10,
though higher values (up to 2^16 -1) are possible
in theory.

The amount of shift will not be lnown in advance --
I will collect the above tuples into a 1 meg array,
figure out the minimum and maximum values, and
do the shifting based on that. Bits cannot fall into
the bitbucket.

TIA

[CBFalconer]

aurgathor wrote:

I use BC 5.02 (long is 32 bit) and I wonder if
there's any efficient code that would allow me
to left shift 64 bit values that I currently use
this way:

typedef struct {
unsigned long lsLong;
unsigned long msLong;
} b64_struct;

Create a mask for the high order bit. This is a separate problem,
to make it clean and portable.

Test the high order bit of the high portion. If set, you have an
overflow, and it is time to barf. If not, shift that portion left.

Test the high order bit of the low portion. If set, increment the
high order portion and reset the bit. Shift that portion left.
done.

This is probably better relegated to a system dependant assembly
language routine, if you can arrange the linkage, and if it is used
enough to be a bind.

--
"If you want to post a followup via groups.google.c om, don't use
the broken "Reply" link at the bottom of the article. Click on
"show options" at the top of the article, then click on the
"Reply" at the bottom of the article headers." - Keith Thompson
More details at: <http://cfaj.freeshell. org/google/>
Also see <http://www.safalra.com/special/googlegroupsrep ly/>

[pete]

aurgathor wrote:

I use BC 5.02 (long is 32 bit) and I wonder if
there's any efficient code that would allow me
to left shift 64 bit values that I currently use
this way:

typedef struct {
unsigned long lsLong;
unsigned long msLong;
} b64_struct;

This struct can be changed if needed.

The least significant part can contain any value
in the 0..2^32 -1 range. The most significant
part; however, will usually be 0, but it may
cintain some smaller values, usually under 10,
though higher values (up to 2^16 -1) are possible
in theory.

The amount of shift will not be lnown in advance --
I will collect the above tuples into a 1 meg array,
figure out the minimum and maximum values, and
do the shifting based on that. Bits cannot fall into
the bitbucket.

msLong is 0x87654321
lsLong is 0x12345678

Left shifted 8:
msLong is 0x65432112
lsLong is 0x34567800

Left shifted 36 more:
msLong is 0x45678000
lsLong is 0x0

/* BEGIN new.c */

#include <stdio.h>

typedef struct {
unsigned long lsLong;
unsigned long msLong;
} b64_struct;

b64_struct leftb64(b64_str uct value, int shift)
{
if (shift > 31) {
shift -= 31;
value = leftb64(value, 31);
value = leftb64(value, shift);
} else {
if (shift > 0) {
value.msLong <<= shift;
value.msLong |= value.lsLong >> 32 - shift;
value.lsLong <<= shift;
}
}
return value;
}

int main(void)
{
b64_struct value = {0x12345678, 0x87654321};

printf("msLong is 0x%lx\n", value.msLong);
printf("lsLong is 0x%lx\n", value.lsLong);
value = leftb64(value, 8);
puts("\nLeft shifted 8:");
printf("msLong is 0x%lx\n", value.msLong);
printf("lsLong is 0x%lx\n", value.lsLong);
puts("\nLeft shifted 36 more:");
value = leftb64(value, 36);
printf("msLong is 0x%lx\n", value.msLong);
printf("lsLong is 0x%lx\n", value.lsLong);
return 0;
}

/* END new.c */
--
pete

[Eric Sosman]

aurgathor wrote:

I use BC 5.02 (long is 32 bit) and I wonder if
there's any efficient code that would allow me
to left shift 64 bit values that I currently use
this way:

typedef struct {
unsigned long lsLong;
unsigned long msLong;
} b64_struct;

Step 0: See whether you can use `unsigned long long' (a
type introduced by C99, but supported as an extension even by
some C90 compilers when operated in lenient modes).

Step 1: Follow the KISS[*] principle, and use

b64_struct x = { ... };
int bits = ...;
x.msLong = (x.msLong << bits) | (x.lsLong >> (32 - bits));
x.lsLong <<= bits;
[*] "Keep It Simple, Stupid!"

Step 2: Try something else. Step 2 is only to be taken if
Steps 0 and 1 don't pan out. Some possible reasons they might not:

.... until and unless this proves to be inadequate. A few
potential reasons for inadequacy:

- Too slow (but you won't know this until you actually
measure it, in the context of the entire program)

- `long long' available only on some but not all of the
systems of interest

- On some systems `long' is wider than 32 bits (easy to
re-spell 32, but you need to decide what to do about
bits 64 and upwards)

- Can only shift by 0 <= bits < 32 (could wrap the above
in a loop if that's a problem)

Step 3: Re-read the subject line. Was it prompted by, er,
any extreme difficulty in getting the code out?

--
Eric Sosman
es*****@acm-dot-org.invalid

[Ben Pfaff]

We'd really appreciate it if you'd keep that kind of language out
of the headers...
--
Ben Pfaff
email: bl*@cs.stanford .edu
web: http://benpfaff.org

[Rod Pemberton]

"Ben Pfaff" <bl*@cs.stanfor d.edu> wrote in message
news:87******** ****@benpfaff.o rg...

We'd really appreciate it if you'd keep that kind of language out
of the headers...

Given that :

1) he didn't use "that kind of language" in his message
2) that f & t use the same finger on QWERTY keyboards
3) it's a common "Freudian slip"

how could you honestly believe it was intentional? I think you rushed to
judgement here.
Rod Pemberton

[ena8t8si]

Rod Pemberton wrote:

"Ben Pfaff" <bl*@cs.stanfor d.edu> wrote in message
news:87******** ****@benpfaff.o rg...

We'd really appreciate it if you'd keep that kind of language out
of the headers...

Given that :

1) he didn't use "that kind of language" in his message
2) that f & t use the same finger on QWERTY keyboards
3) it's a common "Freudian slip"

how could you honestly believe it was intentional? I think you rushed to
judgement here.

Just a quick hint here - try looking up the word "humor".

[Rod Pemberton]

<en******@yahoo .com> wrote in message
news:11******** *************@j 33g2000cwa.goog legroups.com...

Rod Pemberton wrote:

"Ben Pfaff" <bl*@cs.stanfor d.edu> wrote in message
news:87******** ****@benpfaff.o rg...

We'd really appreciate it if you'd keep that kind of language out
of the headers...

Given that :

1) he didn't use "that kind of language" in his message
2) that f & t use the same finger on QWERTY keyboards
3) it's a common "Freudian slip"

how could you honestly believe it was intentional? I think you rushed to judgement here.

Just a quick hint here - try looking up the word "humor".

There was nothing humorous in what he said and there was nothing, such as a
smiley, to indicate that it was humor. I also assumed that since he didn't
post a response to the OP's question, he was offended. That seems to be the
usual case with Pfaff. I also assumed that since he was posting from a
university email, that his school is filtering, flaging, and reviewing his
email for foul language. Do you think my assumptions are invalid?
Rod Pemberton

[Joe Wright]

Rod Pemberton wrote:

<en******@yahoo .com> wrote in message
news:11******** *************@j 33g2000cwa.goog legroups.com...

Rod Pemberton wrote:

"Ben Pfaff" <bl*@cs.stanfor d.edu> wrote in message
news:87******** ****@benpfaff.o rg...
We'd really appreciate it if you'd keep that kind of language out
of the headers...
Given that :

1) he didn't use "that kind of language" in his message
2) that f & t use the same finger on QWERTY keyboards
3) it's a common "Freudian slip"

how could you honestly believe it was intentional? I think you rushed to judgement here.

Just a quick hint here - try looking up the word "humor".

There was nothing humorous in what he said and there was nothing, such as a
smiley, to indicate that it was humor. I also assumed that since he didn't
post a response to the OP's question, he was offended. That seems to be the
usual case with Pfaff. I also assumed that since he was posting from a
university email, that his school is filtering, flaging, and reviewing his
email for foul language. Do you think my assumptions are invalid?
Rod Pemberton

Rod,

Ben Pfaff is well known here and is a friend. He is a graduate student
in a CS PhD program at Stanford University at Palo Alto. I bet he knows
how to spell 'flagging'.

We've known Ben for years. You have come to us only recently. I
personally don't care what assumptions you draw about Ben but you might
consider keeping the negative ones to yourself.

I find your posts to other groups, c.o.m.djgpp for example, helpful and
friendly. Here in comp.lang.c you are a PITA. What gives?

--
Joe Wright
"Everything should be made as simple as possible, but not simpler."
--- Albert Einstein ---

[Rod Pemberton]

"Joe Wright" <jo********@com cast.net> wrote in message
news:MO******** ************@co mcast.com...

Rod Pemberton wrote:

<en******@yahoo .com> wrote in message
news:11******** *************@j 33g2000cwa.goog legroups.com...

Rod Pemberton wrote:
"Ben Pfaff" <bl*@cs.stanfor d.edu> wrote in message
news:87******** ****@benpfaff.o rg...
> We'd really appreciate it if you'd keep that kind of language out
> of the headers...
Given that :

1) he didn't use "that kind of language" in his message
2) that f & t use the same finger on QWERTY keyboards
3) it's a common "Freudian slip"

how could you honestly believe it was intentional? I think you rushed to

judgement here.
Just a quick hint here - try looking up the word "humor".

There was nothing humorous in what he said and there was nothing, such as a smiley, to indicate that it was humor. I also assumed that since he didn't post a response to the OP's question, he was offended. That seems to be the usual case with Pfaff. I also assumed that since he was posting from a
university email, that his school is filtering, flaging, and reviewing his email for foul language. Do you think my assumptions are invalid?

Ben Pfaff is well known here and is a friend. He is a graduate student
in a CS PhD program at Stanford University at Palo Alto. I bet he knows
how to spell 'flagging'.

I don't use a spell checker. You're likely to find occasional misspellings,
rare or alternate spellings, and maybe even a antonym or two in my posts.
We've known Ben for years. You have come to us only recently.
Recently? No. Whether my posts made it to Google or only a few select
servers is a different issue...
I personally don't care what assumptions you draw about Ben but you might
consider keeping the negative ones to yourself.
Did I make a negative comment about Pfaff? I don't see one. Nor, do I see
an untrue one. I see two possible non-negative assumptions. It appears to
me he rushed to judgement. There has been no further clarification from
him.
I find your posts to other groups, c.o.m.djgpp for example, helpful and
friendly. Here in comp.lang.c you are a PITA. What gives?

There are a large number of assholes, frauds, and charlatans here. I'm
dealing with them appropriately. However, in this thread, as far as I can
tell, I've done none of that.
Rod Pemberton