> I know there is a compiler option to let the compiler generate a
> expanded code. The code is still kind of Fortran but not assembly
> file. And we can see loop unrolls and inlined function calls in this
> expanded code. Basically we know where the optimization happens with
> it.
>
> But I can not remeber what this option is. I tried to look into the
> IVF compiler documentation but can not find one.
>
> Anybody can give me a quick answer?
>
> Many thanks!

> On 2008-08-04 18:59:08 -0300, yaqi gmail.com> said:
>

>> I know there is a compiler option to let the compiler generate a
>> expanded code. The code is still kind of Fortran but not assembly
>> file. And we can see loop unrolls and inlined function calls in this
>> expanded code. Basically we know where the optimization happens with
>> it.

>

>> But I can not remeber what this option is. I tried to look into the
>> IVF compiler documentation but can not find one.

>

>> Anybody can give me a quick answer?

>

>> Many thanks!

>
> Perhaps it was a different vendor as not all compilers have the same
> options.
>
> At one time there was a tool that did this sort of thing for Fortran 66. ...

> At one time there was a tool that did this sort of thing for Fortran 66.
> But that was quite a while ago. I do not even recall a name for the tool
> although my impression it was directed as use with IBM 360/91 which are
> truly ancient history anymore. The tool was written in F66, read a F66
> program and output a modified F66 program.

> ...
> At one time there was a tool that did this sort of thing for Fortran 66.
> But that was quite a while ago. I do not even recall a name for the tool
> although my impression it was directed as use with IBM 360/91 which are
> truly ancient history anymore. The tool was written in F66, read a F66
> program and output a modified F66 program.

> Gordon Sande wrote:

>> ...
>> At one time there was a tool that did this sort of thing for Fortran 66.
>> But that was quite a while ago. I do not even recall a name for the tool
>> although my impression it was directed as use with IBM 360/91 which are
>> truly ancient history anymore. The tool was written in F66, read a F66
>> program and output a modified F66 program.

>
> Yup, see:
>
> http://comjnl.oxfordjournals.org/cgi/content/abstract/16/4/322
>
> About 30 years ago, I read the article, called Dr Schneck, and one of
> his minions sent me a tape of it. I may still have a copy of it on
> 9-track tape. (Carefully written at 1600 bpi in order to always be
> readable...)
>
> Besides what was described in the article, it also had the beginnings
> of code to turn Illiac-style array syntax into Fortran DO loops. ...

>> his minions sent me a tape of it. I may still have a copy of it on
>> 9-track tape. (Carefully written at 1600 bpi in order to always be
>> readable...)
>>
>> Besides what was described in the article, it also had the beginnings
>> of code to turn Illiac-style array syntax into Fortran DO loops.
>> This was several years before tools like VAST and KAP.
>>
>> The code was pretty IBM-centric, and assumed one had a machine with
>> 16mbytes of address space. Since I was not using machines with VM
>> at the time, I never really got it running. And by the time I did
>> have a larger machine to use, compiler technology had passed it by.
>>
>> W.

>Er, I thought that too, but the Earth's field changes and is
>equivalent to a moving magnet near your tapes. I lost my 1970's BMD
>Fortran IV source tapes that way, and other tape stuff I wote in
>the1960's...
>
>Tapes get wiped by simple time lapse (as well as self-wiping due to ...