On Thu, 2023-12-14 at 20:50 -0500, James K. Lowden wrote:
> On Fri, 15 Dec 2023 14:43:22 -0500
> "David H. Lynch Jr. via Gcc" <gcc@gcc.gnu.org> wrote:
> > Right now I am just focused on some means to deliver support. 
> Hi David, 
> My colleague Bob Dubner and I have been extending GCC every day for
> the last two years.  I wonder if we might be of some use to you.  
> I only faintly hope our project can benefit from your work. We're
> adding a Cobol front end to GCC.  Cobol has built-in sort functions,
> both on disk and in memory, and a rich data-description language.

I have not thought regarding COBOL.  I have not used COBOL since the
early 80's.  
Regardless, The work we are doing likely has a high synergy with COBOL.
The core issue would be that COBOL is not a high profile target. 

The current work is focused on delivering a demonstration project with
enough sex appeal to attract the resources to more fully develop our

This project started with a major memory manufacturer about 5 years
ago.  Dr. Trout and I developed it there. We secured the rights to our
work when we left.  We have been working on it on our own since. 

The original "target" was Sort-in-memory - which we demonstrated on
FPGA's several years ago. We were very close to having DRAM fabricated
when we left. We have subsequently expanded our idea to incorporate
numerous memory access/addressing options. 

One of the target applications is handling of sparse arrays, and
directly accessing the cells using existing sparse array notation. 

That is particularly applicable to AI. 

We are in the process of trying to take a standard AI reference
application - recognizing handwritten numbers and being able to
demonstrate everything - from the application through the compiler to
the memory access.  Doing a complete end to end problem helps get the
mundane issues that have not been addressed completed. 

The AI app we are using is in C/C++ so that is our focus. 

Our long term objective is the hardware itself - currently our target
is DRAM, but our work is applicable to all forms of storage - from DDR
through to Hard disks. 

Frankly there are applications we have not even thought of. 

Technically this is Processor-In-Memory - which others such as Samsung
are doing.  But our approach is quite different, we are dual purposing
the addressing logic and making use of a 4 bit processor to dynamically
alter the way memory is addressed/accessed. 

Right now MY priority is adding support to a c/c++ compiler for our
memory. I can  do this myself, But I have not worked on compilers for
30 years and I have not ever worked on GCC. 

I am looking for any help I can get - pointers as to where to start
with GCC, docs or howto's through to someone that wishes to participate
in the project.  There is a potential for compensation - we are seeking
a grant, though our long term goals are partnership with a Memory

We have worked out a syntax with out preprocessor, but we are not
committed to that. 

There are some aspects that we are not sure how map into a programming
language.  Regardless the goal is not to get everything right it is to
prove the value of the concept. 

Currently AI is where all the focus is - which is why we are doing an
end to end AI application. Everything at the MEMSYS conference where we
presented our work was either directly or indirectly about AI. 

Absolutely COBOL with Sort in Memory would be wonderful, but it is not
likely to attract interest int eh addressing technology that is out key

Regardless, anyone interested can contact me either on this list or
directly via email.  I am looking for help and guidance and I do not
have  a preconception for what that might mean. 


> There is more potential there than might seem at first blush, and I
> would welcome the opportunity to explain in detail if you're
> interested.  
> If your objective is simply to extend C to support content
> addressable
> memory, then we might still be of some help.  I don't know anything,
> really, about the C front-end, but Bob has experience getting
> Generic to generate code.  He might be able to answer some of your
> questions, if nothing else.
> Let me know what you think.  
> Kind regards, 
> --jkl

Reply via email to