CGEN output files

CGEN output files

[Carl Frankel]

Question:  Does CGEN produce .l/.y files for the assembler
           as a part of its output?  If not now, is this
           planned, and if so, in what time frame, more or
           less.

I ask this as a cpu verifier.  I need to generate a 
directed/random instruction sequence generator, and
having .l/.y files to augment will be a tremendous
time savings in creating the sequence generator.

Thanks,

Carl Frankel


=======================================================================
+ Carl B. Frankel, Ph.D.                    Ubicom                    +
+ Manager, ASIC Design Verification         1330 Charleston Road      +
+                                           Mountain View, CA 94043   +
+ carl@ubicom.com   http://www.ubibom.com   (Corp) 650-210-1500       +
=======================================================================
+ (Direct) 650-210-1537     (Cell) 650-740-1565    (Fax) 650-210-8715 +
=======================================================================
+ What's Punishing Gets Priority         What's Measured Gets Managed +
+ What's Rewarded Gets Repeated          What's Noticed Gets Narrated +
=======================================================================

Re: CGEN output files

[Frank Ch. Eigler]

Hi -

On Fri, Mar 09, 2001 at 04:16:14PM -0800, Carl Frankel wrote:
: Question:  Does CGEN produce .l/.y files for the assembler
:            as a part of its output?  [...]

Do you mean lex/yacc input files?  No.  CGEN-based assemblers
use a less formal run-time parser system built into the opcodes
library and to a lesser extent gas.


: I ask this as a cpu verifier.  I need to generate a 
: directed/random instruction sequence generator  [...]

You may find the existing assembler test case generator
of use.  See cgen's gas-test module and Makefile target.
It is fairly simpleminded (especially with regard to the
lack of diversity in the operand values), but with perhaps
not much Scheme code, it could be made to emit a stream of
randomized instructions.  


- FChE
-- 
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.4 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE6qZlzVZbdDOm/ZT0RApS1AJ9JTC4AIhp5RIjl6BZwlsWAUyM+EQCfXAYQ
18oS5iFPbp/vHlvJ2COXUSo=
=n9IG
-----END PGP SIGNATURE-----

Re: CGEN output files

[Doug Evans]

Frank Ch. Eigler writes:
 > On Fri, Mar 09, 2001 at 04:16:14PM -0800, Carl Frankel wrote:
 > : Question:  Does CGEN produce .l/.y files for the assembler
 > :            as a part of its output?  [...]
 > 
 > Do you mean lex/yacc input files?  No.  CGEN-based assemblers
 > use a less formal run-time parser system built into the opcodes
 > library and to a lesser extent gas.

Creating .l/.y files wasn't needed at the time (and isn't needed
for the typical risc architecture), though isn't precluded as a useful
addition.

CGEN output files

[Doug Evans]

Carl  Frankel writes:
 > 
 > Question:  Does CGEN produce .l/.y files for the assembler
 >            as a part of its output?  If not now, is this
 >            planned, and if so, in what time frame, more or
 >            less.
 > 
 > I ask this as a cpu verifier.  I need to generate a 
 > directed/random instruction sequence generator, and
 > having .l/.y files to augment will be a tremendous
 > time savings in creating the sequence generator.

It should be straightforward to take the existing testcase generator
and enhance it to suit such purposes.
[assuming you know Scheme, but it's a joy to program in anyway]