From Newsgroup: comp.lang.fortran

On Thu, 17 Apr 2025 12:24:22 +0200, R Daneel Olivaw wrote:

Steven G. Kargl wrote:

On Wed, 16 Apr 2025 07:44:08 +0000, Lawrence D'Oliveiro wrote:

On Tue, 15 Apr 2025 18:28:31 -0500, Lynn McGuire wrote:

On 4/15/2025 6:14 PM, Lawrence D'Oliveiro wrote:

On Mon, 14 Apr 2025 23:50:32 -0500, Lynn McGuire wrote:

Got rid of a few nasty bugs like:

iword = 6Habcdef

Surely whether thatrCOs a bug or not would depend on the type of rCLiwordrCY
...

iword is a implicit 4 byte integer capable of storing 4 characters.

I thought you got rid of all the implicit typing.

Implicit typing has nothing do with numeric storage size.

program foo
use iso_fortran_env, only : numeric_storage_size
integer :: j = 0
i = 6Habcdef ! i has an implicit type of default integer kind
j = 6Habcdef ! j has an explicit type of default integer kind
print *, i, j
print *, numeric_storage_size
end program foo

(snip)

I believe you'll find that some of that syntax did not exist under
Fortran 77, in particular the "use" line.
Did Lynn convert to F90 first?

Of course, the USE statement was not part of FORTRAN 77.
(It was added in Fortran 90). The intrinsic module
iso_fortran_env was not part of FORTRAN 77. (It was
added in Fortran 2003). I included that line to easily
get to the number of bits in a numeric storage unit.
From ANSI X3.9-1978 (i.e., FORTRAN 77):

An integer, real, or logical datum has one numeric
storage unit in a storage sequence.
...
This standard does not specify a relationship between
the storage sequence concept and the physical properties
or implementation of storage.

The size of the unit was not specified in FORTRAN 77, and
this concept is still alive in Fortran 2023.
--
steve
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

In comp.lang.fortran Lynn McGuire <lynnmcguire5@gmail.com> wrote:

On 11/23/2022 5:36 PM, Tran Quoc Viet wrote:

On Saturday, November 19, 2022 at 1:01:25 PM UTC+7, Lynn McGuire wrote:

We are converting a 700,000+ Fortran 77 lines of code plus 50,000+ C++
lines of code engineering software product to C++. With all that code,

Mentioned recently on Hacker News:

https://www.lanl.gov/media/publications/1663/0125-llm-translation

"...O???Malley is taking open-source LLMs, running them on Lab computers, and plying the models with a technique called retrieval-augmented generation
(RAG), where generative models are enhanced with data from external
sources. The idea is to train the models to translate from Fortran to C++
using what???s known as few-shot learning..."
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

On 4/21/2025 11:53 PM, David Duffy wrote:

In comp.lang.fortran Lynn McGuire <lynnmcguire5@gmail.com> wrote:

On 11/23/2022 5:36 PM, Tran Quoc Viet wrote:

On Saturday, November 19, 2022 at 1:01:25 PM UTC+7, Lynn McGuire wrote: >>>> We are converting a 700,000+ Fortran 77 lines of code plus 50,000+ C++ >>>> lines of code engineering software product to C++. With all that code,

Mentioned recently on Hacker News:

https://www.lanl.gov/media/publications/1663/0125-llm-translation

"...O???Malley is taking open-source LLMs, running them on Lab computers, and plying the models with a technique called retrieval-augmented generation (RAG), where generative models are enhanced with data from external
sources. The idea is to train the models to translate from Fortran to C++ using what???s known as few-shot learning..."

Cool, I am not the only one !

Lynn


--- Synchronet 3.21a-Linux NewsLink 1.2