cmucl vs. gcl



Hi,
  I started maxima again and just loaded the matrix stuff with cmucl.

Things now work--Thanks for the replies.

Apparently, I had a bunch of stuff loaded in memory from other work
when I tried the Esparse program before.

-sen

On Mon, 11 Dec 2006, sen1 at math.msu.edu wrote:

> On Mon, 11 Dec 2006, Raymond Toy wrote:
>
>> sen1 at math.msu.edu wrote:
>>> On Mon, 11 Dec 2006, sen1 at math.msu.edu wrote:
>>>
>>> whoops, in the last post I had some other aliases (e.g. "flt" for "float",
>>>  "prt" for "print".
>>>
>> You missed one flt and one prt.
>>
>> How am I supposed to run this?  Esparse(A,27,30,1e-10)?
>>
>> If so, I have no problems running it on my copy of cmucl (on solaris).
>> However, Esparse just says I should increase MAX or decrease ERR.
>
> The program is designed for an irreducible non-negative matrix.
>
> Such matrices have a unique real eigenvalue of largest magnitude.
>
> If A is such a matrix, the program takes a random vector v and
>
> successively does the following.
>
> 1. compute the product A.v
> 2. normalize (i.e., divide by the norm) and repeat until either
>    MAX iterates have been run or the difference between the current
>    unit vector and the previous is less than ERR.
>
> When ERR is achieved, the next length of A.v is approximately the
> largest eigenvalue.
>
> If the message is printed, just rerun the program with MAX larger
> (ie., more iterates), or ERR bigger (not a good idea if you really
> want an estimate for the largest eigenvalue).
>
> So, you managed to run it with the 27x27 C on cmucl?
>
> With what MAX and what ERR?
>
> With gcl, here is what I get with various MAX and ERR values.
>
> (%i12) showtime: true;
> Evaluation took 0.00 seconds (0.00 elapsed) using 96 bytes.
> (%o12)                               true
> (%i13) Esparse(A,27,30,1e-10);
> MAX reached and ERR not achieved; try decreasing
>                    ERR or increasing MAX
> Evaluation took 2.16 seconds (2.25 elapsed) using 9.783 MB.
> (%o13)                               done
> (%i14) Esparse(A,27,130,1e-10);
> Evaluation took 8.20 seconds (14.09 elapsed) using 33.274 MB.
> (%o14)                          1.5277485879658
>
>
> So, the eigenvalue is approx 1.5277...
>
> Using matlab, one gets:      1.52774858799907,
>
> so the program seems to work (although it is a bit slow).
>
>
>>
>>
>> Ray
>>
>
>

-- 
  ---------------------------------------------------------------------------
  | Sheldon E. Newhouse            |    e-mail: sen1 at math.msu.edu           |
  | Mathematics Department         |       				   |
  | Michigan State University      | telephone: 517-355-9684                |
  | E. Lansing, MI 48824-1027 USA  |       FAX: 517-432-1562                |
  ---------------------------------------------------------------------------