[comp.sys.transputer] C compilers and loaders

K312240@aearn (Klaus Kusche) (03/05/88)

     
     
Second try, the first one did not come back to me.
     
     
Dear Mailing List!
     
There were some messages about C Compilers, loaders etc. on the mailing
list during the last few weeks. In the following, I will shortly
describe what I know about it. If anyone out there knows more, or if I
missed your product, please *tell* me, I'm very interested in transputer
C compilers, too!.
     
1.) Inmos C, Lattice Logic (3L) C:
Both are essentially the same. New releases have just been sent out (I
don't have mine yet). Not very comfortable, far from ANSI (both language
and library), but available via inmos distributors for a *wide* range of
transputer systems, and via many other companies, too.
     
Parallel facilities: I heard that Occam-like facilities will be included
in the new release. Together with the new stand-alone Occam compiler
package, it will be possible to write mixed-mode (Occam, C, Fortran,
Pascal) parallel programs without that clumsy importing of
alien-language program parts into TDS --- Occam no longer necessary!
     
The linker, configurer and loader included in that compiler
will offer similar facilities as the TDS (even for multiproc's I heard),
but are completely independent of it.
     
I think these are the products Adrian Cockcroft referred to in his last
mail.
     
2.) Lattice Logic (3L) Parallel C:
I heard that this is just 3L C with some parts of the standalone Occam
system mentioned above --- should offer most Occam and TDS facilities
for C freaks (i.e. without any Occam at all).
     
3.) Dwight's (Pentasoft's) C:
As far as I heard, the most complete and highest-quality C compiler
available (full ANSI, optimizing code generator,...). Oriented towards
UNIX environments, available as a cross-compiler running on Suns,... .
Used at Cornell.
     
I don't know anything about its parallel features or any other details,
but I would like to receive more information about this product.
     
Hence, if anyone knows Dwight VandenBerghe's e-mail address,
*p l e a s e   t e l l* me.
     
4.) Bailey's (Logical Sytem's) C:
A C compiler for use in PC environments (B004), but sources are said to
be included, and it is said to be easily portable. Approaches ANSI (both
language and library), offers quite good parallel facilities (more than
Occam), and comes with multi-processor loaders and similar tools.
     
5.) Definicon C:
I don't know much about it (again, if anyone has an e-mail address of
that company or person, or uses this product, please *tell* me),
but I heard that its facilities are similar to Bailey's C
(also including multi-transputer support).
     
6.) Other Products:
There seems to be quite a variety of other products, especially in the
UK. I think most of them are based on Inmos /3L C, but I'm not sure.
     
About Carsten Petersen's comments: I would *not* port anything from C to
Occam: There is at least as much C development as Occam development done
on transputers, and in my mind, Occam is quite a primitive, inconvenient
language (I better do not give any comments about the TDS...).
     
Klaus (K312240@AEARN.BITNET).

K312240@aearn (Klaus Kusche) (03/10/88)

     
     
Dear Mailing List!
     
There were some messages about C Compilers, loaders etc. on the mailing
list during the last few weeks. In the following, I will shortly
describe what I know about it. If anyone out there knows more, or if I
missed your product, please *tell* me, I'm very interested in transputer
C compilers, too!.
     
1.) Inmos C, Lattice Logic (3L) C:
Both are essentially the same. New releases have just been sent out (I
don't have mine yet). Not very comfortable, far from ANSI (both language
and library), but available via inmos distributors for a *wide* range of
transputer systems, and via many other companies, too.
     
Parallel facilities: I heard that Occam-like facilities will be included
in the new release. Together with the new stand-alone Occam compiler
package, it will be possible to write mixed-mode (Occam, C, Fortran,
Pascal) parallel programs without that clumsy importing of
alien-language program parts into TDS --- Occam no longer necessary!
     
The linker, configurer and loader included in that compiler
will offer similar facilities as the TDS (even for multiproc's I heard),
but are completely independent of it.
     
I think these are the products Adrian Cockcroft referred to in his last
mail.
     
2.) Lattice Logic (3L) Parallel C:
I heard that this is just 3L C with some parts of the standalone Occam
system mentioned above --- should offer most Occam and TDS facilities
for C freaks (i.e. without any Occam at all).
     
3.) Dwight's (Pentasoft's) C:
As far as I heard, the most complete and highest-quality C compiler
available (full ANSI, optimizing code generator,...). Oriented towards
UNIX environments, available as a cross-compiler running on Suns,... .
Used at Cornell.
     
I don't know anything about its parallel features or any other details,
but I would like to receive more information about this product.
     
Hence, if anyone knows Dwight VandenBerghe's e-mail address,
*p l e a s e   t e l l* me.
     
4.) Bailey's (Logical Sytem's) C:
A C compiler for use in PC environments (B004), but sources are said to
be included, and it is said to be easily portable. Approaches ANSI (both
language and library), offers quite good parallel facilities (more than
Occam), and comes with multi-processor loaders and similar tools.
     
5.) Definicon C:
I don't know much about it (again, if anyone has an e-mail address of
that company or person, or uses this product, please *tell* me),
but I heard that its facilities are similar to Bailey's C
(also including multi-transputer support).
     
6.) Other Products:
There seems to be quite a variety of other products, especially in the
UK. I think most of them are based on Inmos /3L C, but I'm not sure.
     
About Carsten Petersen's comments: I would *not* port anything from C to
Occam: There is at least as much C development as Occam development done
on transputers, and in my mind, Occam is quite a primitive, inconvenient
language (I better do not give any comments about the TDS...).
     
Klaus (K312240@AEARN.BITNET).