Re: a big limit of mathematica?
- From: LumisROB <lumisrob_NOGOOD_@xxxxxxxxx>
- Date: Mon, 12 Sep 2005 22:23:06 GMT
On 12 Sep 2005 11:52:08 -0700, "Daniel Lichtblau" <danl@xxxxxxxxxxx>
wrote:
>
>
>In[2]:= test[8]
>5
>{0.001, 0, 0, 0.001}
>6
>{0.011998, 0, 0, 0.020997}
>7
>{0.088987, 0, 0, 0.125981}
>8
>{0.679896, 0, 0, 1.12283}
>
>This reevaluation semantics used in Mathematica is admittedly not
>always optimal. But it is also, in this example, not so bad as one
>might guess based only on knowing the first time behavior.
>
>Daniel Lichtblau
>Wolfram Research
Ok I have understood this concept
This behavior is predictable knowing as mathematica kernel is
implemented
The true problem is:
is it possible to foresee in advance having a computer,
an OS (32 or 64 Bit) and Mathematica(32 or 64 Bit) which dimensions of
a determined problem I succeed in resolving?
With Matlab or Maple or C I have always succeeded in banal way in
making this calculation (I work on problems treated with Finite
Element method).
With Mathematica no and anybody me ago a practical example
For instance because he doesn't succeed in calculating
m = Table[(i - j) + 3./j^4 + 50*i, {i, 1, 10^5}, {j, 1, 10^5}];
m[[1]].m[[1]]
ByteCount[1.0]=16
if I go to calculate this matrix
logic would like 10^5 * 10^5 *16 Bytes * 10^-9 = 160 GB
This problem on my hardware is interrupted for lack of memory (as
usual).
With Matlab or Maple on Same Hardware it is resolved (Ok!! with
exasperated slowness..... but they resolve )
Matlab (or a compiled program) performs this calculation because, I
beklieve, the data that have to contemporarily enter
the memory don't exceed the available Ram and the Virtual memory
handles the rest
Because nobody responds to this question?
is it perhaps the error that I commit so banal not to even receive an
answer?
Cheers
Roberto
.
- Follow-Ups:
- Re: a big limit of mathematica?
- From: Daniel Lichtblau
- Re: a big limit of mathematica?
- References:
- a big limit of mathematica?
- From: LumisROB
- Re: a big limit of mathematica?
- From: Peter Pein
- Re: a big limit of mathematica?
- From: LumisROB
- Re: a big limit of mathematica?
- From: Richard Fateman
- Re: a big limit of mathematica?
- From: carlos
- Re: a big limit of mathematica?
- From: Richard Fateman
- Re: a big limit of mathematica?
- From: carlos
- Re: a big limit of mathematica?
- From: Richard J. Fateman
- Re: a big limit of mathematica?
- From: Daniel Lichtblau
- a big limit of mathematica?
- Prev by Date: Re: remarkable results in numericals
- Next by Date: Re: linear system stability
- Previous by thread: Re: a big limit of mathematica?
- Next by thread: Re: a big limit of mathematica?
- Index(es):
Relevant Pages
|