On Sat, 23 Dec 2006, Barton Willis wrote:
> For big 'm', your calculation suffers from subtractive
> cancelation; try bfloat instead of float:
>
> (%i11) huge : gain(61,%i*2*%pi*60)$
> (%i12) bfloat(huge), fpprec : 123;
> (%o12) 1.1462233314862319837372937172[66
> digits]7101238230788110302499429193b0
> (%i13) float(huge);
> (%o13) 0.0
>
> Barton
thanks a lot, this solves the mystery and helps me further
best regards,
hugo
>
> -----maxima-bounces at math.utexas.edu wrote: -----
>
>> To: Barton Willis
>> From: Hugo Coolens
>> Sent by: maxima-bounces at math.utexas.edu
>> Date: 12/23/2006 02:43PM
>> cc: maxima at math.utexas.edu
>> Subject: Re: [Maxima] newbie question, forcing
>> numeric result
>>
>>
>>
>>
>> On Sat, 23 Dec 2006, Barton Willis wrote:
>>
>>> Try *something* like
>>>
>>> (%i1) mypow(a,b) := if a=0 and b= 0 then 1 else
>> a^b;
>>> paldn(m,s) :=
>>>
>> (sum(binomial(2*m-i,i)*mypow(s,i),i,0,m-1))/(sum(
>> binomial(2*m-i,i)*mypow(s,i),i,0,m));
>>>
>>> gain(m,s) := abs(paldn(m,s));
>>> pairs: makelist([0.1*omega,
>> gain(2,%i*omega)],omega,0,5);
>>>
>>> (%o1) mypow(a,b):=if a=0 and b=0 then 1 else
>> a^b
>>> (%i2)
>>> (%o2)
>>>
>> paldn(m,s):=sum(binom(2*m-i,i)*mypow(s,i),i,0,m-1
>> )/sum(binom(2*m-i,i)*mypow(s,i),i,0,m)
>>>
>>> (%i3)
>>> (%o3) gain(m,s):=abs(paldn(m,s))
>>> (%i4)
>>> (%o4)
>>>
>> [[0,1],[0.1,sqrt(10)/3],[0.2,sqrt(37)/(3*sqrt(5))
>> ],[0.3,sqrt(82)/sqrt(145)],[0.4,sqrt(145)/(3*sqrt
>> (41))],[0.5,sqrt(226)/(3*sqrt(89))]]
>>>
>>> (%i5)
>>>
>>> Notes:
>> thanks for the instructive reply, strange enough
>> I am confronted now with
>> another problem. The above defined function
>> gain(m,s) stops functioning
>> correctly when m>=60 as you can see below. It is
>> a complete mystery to me
>> why this happens:
>>
>> (%i19) float(gain(59,%i*2*%pi*60));
>>
>> (%o19) 1.14706260261368
>>
>> (%i18) float(gain(60,%i*2*%pi*60));
>>
>> (%o18) 0.0
>>
>> (%i20) float(gain(61,%i*2*%pi*60));
>>
>> (%o20) 0.0
>> _______________________________________________
>> Maxima mailing list
>> Maxima at math.utexas.edu
>> http://www.math.utexas.edu/mailman/listinfo/maxim
>> a
>