A tool that suggests optimized logic for a piece ofcode/module/function

[karthikbalaguru]

Hi,
There are certain editors that highlight
the syntax/color for datatypes/variables
or comments etc.

Similarly,
Is there a tool for C language that
could suggest an optimized/alternate
programming logic for the function that
is written ?

The optimized/alternate logic can be
suggested as soon as we finish coding
for one function or it can be suggested
as soon as the code is compiled/parsed
by that tool.

It will be even more helpful if that tool
also provides the cycle counts, cache
usage, cache misses and lines of code
also.

It would be better if that tool has an
option to enable / disable this feature
either through compile time or some
other configurations.

Any ideas ?

Thx in advans,
Karthik Balaguru

 

[acd]

Hi,
There are certain editors that highlight
the syntax/color for datatypes/variables
or comments etc.

Similarly,
Is there a tool for C language that
could suggest an optimized/alternate
programming logic for the function that
is written ?

The optimized/alternate logic can be
suggested as soon as we finish coding
for one function or it can be suggested
as soon as the code is compiled/parsed
by that tool.

It will be even more helpful if that tool
also provides the cycle counts, cache
usage, cache misses and lines of code
also.

It would be better if that tool has an
option to enable / disable this feature
either through compile time or some
other configurations.

Any ideas ?

Thx in advans,
Karthik Balaguru

Brain?

Seriously, I think this is impossible.
What I can recommend to you are some good books, including
(More) Programming Perls
and
Hacker's delight.

Andreas

 

[scattered]

Brain?

Seriously, I think this is impossible.
What I can recommend to you are some good books, including
(More) Programming Perls
and
Hacker's delight.

Andreas- Hide quoted text -

- Show quoted text -

I don't see why it would be impossible. Optimizing compilers exist and
by using similar techniques it should be possible to write a compiler
that uses C as both the source and target language. How useful this
would be is another question. The resulting code would need to be
humanly readable if the tool would be of any use and the would
probably place severe restrictions on the sorts of optimizations that
can be done. I would be surprised if no work along these lines has
been done.

-scattered

 

[Andrew Poelstra]

I don't see why it would be impossible. Optimizing compilers exist and
by using similar techniques it should be possible to write a compiler
that uses C as both the source and target language. How useful this
would be is another question. The resulting code would need to be
humanly readable if the tool would be of any use and the would
probably place severe restrictions on the sorts of optimizations that
can be done. I would be surprised if no work along these lines has
been done.

Indeed, some of what he suggested (loop counting, etc) would be not
only very useful, but would be an excellent parsing/analysis learning
experience if the OP were to try and implement it himself.

And C would be an excellent language for this, because it has so few
keywords and such straightforward branch control contructs (function
pointers aside).

 

[Walter Banks]

I don't see why it would be impossible. Optimizing compilers exist and
by using similar techniques it should be possible to write a compiler
that uses C as both the source and target language. How useful this
would be is another question. The resulting code would need to be
humanly readable if the tool would be of any use and the would
probably place severe restrictions on the sorts of optimizations that
can be done. I would be surprised if no work along these lines has
been done.[/QUOTE]

Programming Pearls and Hacker's delight are both must haves
even if you are only remotely interested programming algorithms.

There has been quite a bit of work done in compilers at the
expression level to optimize algorithms and statements
implementing functionally equivalent statements in the
generated code.

The problem in the bigger scale suggested by the op is the
ability to recognize the larger overview of the programming
objective at the scale of turning a bubble sort into a quick
sort.

At the statement level recognizing functionality is generally
a simple task. At the function level this a significantly larger
problem.

The solution may be mostly hard work and processing
cycles. A start may be to catalog common functions
and implementations and create a significant tool
to recognize these functions from source code extracting
out key function parameters and then using these to select
alternative implementations.

As well as the data base of common functions a significant
piece of expert system software would need to be written
to extract knowledge out of the source.

Do-able probably.

Regards,

 

[Walter Banks]

Is there a tool for C language that
could suggest an optimized/alternate
programming logic for the function that
is written ?

The optimized/alternate logic can be
suggested as soon as we finish coding
for one function or it can be suggested
as soon as the code is compiled/parsed
by that tool.

It will be even more helpful if that tool
also provides the cycle counts, cache
usage, cache misses and lines of code
also.

It would be better if that tool has an
option to enable / disable this feature
either through compile time or some
other configurations.

Metrics are available in many (especially embedded system)
compilers as part of the listing and report files. Better
compilers do a good job of instruction scheduling and
cache management in their code generation.

The issue is the larger one of function rewriting where
the problem is larger and less work has been done
looking for solutions.


Regards,

 

[Andy]

I don't see why it would be impossible. Optimizing compilers exist and
by using similar techniques it should be possible to write a compiler
that uses C as both the source and target language. How useful this
would be is another question. The resulting code would need to be
humanly readable if the tool would be of any use and the would
probably place severe restrictions on the sorts of optimizations that
can be done. I would be surprised if no work along these lines has
been done.

The optimizations would have to be limited to that subset that can be
implemented at the C level. That might be difficult to do, since the
optimization is not usually concerned with whether or not the
optimized version can be portrayed in C.

The generated C code would have to look similar (same variable names,
etc.) to the original to be useful.

And if the compiler can show you those optimizations in the form of
modified code, why not have it skip showing you the code and just
optimize it behind the scenes like it always does?

Andy

 

[karthikbalaguru]

Metrics are available in many (especially embedded system)
compilers as part of the listing and report files. Better
compilers do a good job of instruction scheduling and
cache management in their code generation.

Yes, many embedded system based
compilers support the report generation
of cache misses, cache usage and
instruction cycles. I would like to
convey that it would be better if the
same tool provides those various
optimization reports for both the
logics apart from the optimized/
alternate logic suggested by it.

The issue is the larger one of function rewriting where
the problem is larger and less work has been done
looking for solutions.

It is strange that no much research or
work is done in this area ? Strange !!

I think, this is one of the most required
tool for efficient software development
in any kind of platform.

Thx in advans,
Karthik Balaguru

 

[karthikbalaguru]

Programming Pearls and Hacker's delight are both must haves
even if you are only remotely interested programming algorithms.

There has been quite a bit of work done in compilers at the
expression level to optimize algorithms and statements
implementing functionally equivalent statements in the
generated code.

True !

The problem in the bigger scale suggested by the op is the
ability to recognize the larger overview of the programming
objective at the scale of turning a bubble sort into a quick
sort.

Exactly, this is the actual problem that
has to be addressed.

At the statement level recognizing functionality is generally
a simple task. At the function level this a significantly larger
problem.

The solution may be mostly hard work and processing
cycles. A start may be to catalog common functions
and implementations and create a significant tool
to recognize these functions from source code extracting
out key function parameters and then using these to select
alternative implementations.

As well as the data base of common functions a significant
piece of expert system software would need to be written
to extract knowledge out of the source.

Do-able probably.

Thx in advans,
Karthik Balaguru

 

[karthikbalaguru]

The optimizations would have to be limited to that subset that can be
implemented at the C level.

True !

That might be difficult to do, since the
optimization is not usually concerned with whether or not the
optimized version can be portrayed in C.

True, but i am looking only at the
C level optimization.

The generated C code would have to look similar (same variable names,
etc.) to the original to be useful.

I think, in such scenarios, the tool should
be having tremendous knowledge about
the software/program and take care
end to end. Maybe, the tool can generate
report on the changes and the reason for
the changes so the developer can use
it for further developments or any other
maintenance activities on that code.

And if the compiler can show you those optimizations in the form of
modified code, why not have it skip showing you the code and just
optimize it behind the scenes like it always does?

Maybe, it would be better if it
can do it before user rather than
behind the scenes so that the user
is aware of the kind of optimzations
that are being done and could be
taken care. It is also needed so that
the developer will be able to take care
of further developments on the same
program/software and also for the
maintenance activities on that
program/software.

Thx in advans,
Karthik Balaguru

 

[David Schwartz]

I don't see why it would be impossible. Optimizing compilers exist and
by using similar techniques it should be possible to write a compiler
that uses C as both the source and target language. How useful this
would be is another question. The resulting code would need to be
humanly readable if the tool would be of any use and the would
probably place severe restrictions on the sorts of optimizations that
can be done. I would be surprised if no work along these lines has
been done.

That would be a useless tool, all it would do would be obfuscate. If
the code contains optimizations that can be made by machine, what is
the point of modifying the source code? Let the machine make the
optimizations when the code is compiled and keep the source intact.

DS

 

[Andrew Poelstra]

That would be a useless tool, all it would do would be obfuscate. If
the code contains optimizations that can be made by machine, what is
the point of modifying the source code? Let the machine make the
optimizations when the code is compiled and keep the source intact.

DS

The idea would be that it could provide suggestions, not necessarily
implement them or even come up with a concrete solution. Even something
as simple as highlighting a function as "O(n^3)" would be helpful.

Saying something like "this looks like a Bubblesort" and linking to a
database (or wiki link) of sorting algorithms would also be nice, even
if the machine didn't understand the data structures or nature of the
data well enough to help.

 

[Jasen Betts]

Hi,
There are certain editors that highlight
the syntax/color for datatypes/variables
or comments etc.

Similarly,
Is there a tool for C language that
could suggest an optimized/alternate
programming logic for the function that
is written ?

The optimized/alternate logic can be
suggested as soon as we finish coding
for one function or it can be suggested
as soon as the code is compiled/parsed
by that tool.

It will be even more helpful if that tool
also provides the cycle counts, cache
usage, cache misses and lines of code
also.

It would be better if that tool has an
option to enable / disable this feature
either through compile time or some
other configurations.

Any ideas ?

you seem to be asking for a compiler that is
tightly bopund to an editor, (which seems contrary to the unix
philosophy) such compiler is likely to optimise worse than GCC.

Tweaked C code is likely to only work well on a single architecture.

optimising compilers often alter the order in which statements are
executed etc...

If you really want to optimise for a single architecture
you should probably be coding in assembler anyway.

Thx in advans,
Karthik Balaguru

--- news://freenews.netfront.net/ - complaints: (e-mail address removed) ---

 

[Stefan Ram]

Similarly,
Is there a tool for C language that
could suggest an optimized/alternate
programming logic for the function that
is written ?

I have not read the whole thread, so, I am sorry,
if this was already mentioned:

http://www.supercompilers.com/white_paper.shtml

 

[Rich Webb]

That would be a useless tool, all it would do would be obfuscate. If
the code contains optimizations that can be made by machine, what is
the point of modifying the source code? Let the machine make the
optimizations when the code is compiled and keep the source intact.

What he's looking for is not compiler optimizations but optimization of
the underlying algorithm. As somebody else mentioned, seeing a bubble
sort and deciding that a quick sort would be more appropriate.

 

[Keith Thompson]

What he's looking for is not compiler optimizations but optimization of
the underlying algorithm. As somebody else mentioned, seeing a bubble
sort and deciding that a quick sort would be more appropriate.

Of course, that particular example is not terribly useful in practice,
since it's easy enough to just not write a bubble sort in the first
place. But it's a good as an example because it's easy to understand.

 

[Ed Prochak]

True !


True, but i am looking only at the
C level optimization.


I think, in such scenarios, the tool should
be having tremendous knowledge about
the software/program and take care
end to end. Maybe, the tool can generate
report on the changes and the reason for
the changes so the developer can use
it for further developments or any other
maintenance activities on that code.


Maybe, it would be better if it
can do it before user rather than
behind the scenes so that the user
is aware of the kind of optimzations
that are being done and could be
taken care. It is also needed so that
the developer will be able to take care
of further developments on the same
program/software and also for the
maintenance activities on that
program/software.

Thx in advans,
Karthik Balaguru

There is such a tool. It is called a code review.

 

[Paul Keinanen]

It is strange that no much research or
work is done in this area ? Strange !!

I think, this is one of the most required
tool for efficient software development
in any kind of platform.

If the problem on average is badly trained programmers, the obvious
solution should be improving the training of programmers.

A tool that points out suboptimal constructs etc. might be useful in
the training phase, but what is the point of using it in actual
program production ?

 

[Chris McDonald]

If the problem on average is badly trained programmers, the obvious
solution should be improving the training of programmers.

A tool that points out suboptimal constructs etc. might be useful in
the training phase, but what is the point of using it in actual
program production ?

If sufficiently sophisticated it could be used as a regression tester
for code submitted to a revision system.

This presumes, however, that it's better to have "optimal" source code
than to have readble/expressive source code.

 

[Phil Carmody]

What he's looking for is not compiler optimizations but optimization of
the underlying algorithm. As somebody else mentioned, seeing a bubble
sort and deciding that a quick sort would be more appropriate.

Unless it was only ever a handful of items, in which case an
insertion sort might be more appropriate. Know your N (most
important when N might be large, of course, but also when it's
small). However, anything which brings about the nuking of
bubblesort in any context is an improvement.

Phil