On Wed, Jul 8, 2009 at 11:13 AM, smichr<[email protected]> wrote:
>
>
>
> On Jul 7, 10:38 am, Ondrej Certik <[email protected]> wrote:
>> On Sat, Jul 4, 2009 at 6:52 PM, smichr<[email protected]> wrote:
>>
>> > I noticed that fastlog in evalf computes an approximation to log2(x).
>> > Is there any reason not to call it  fastlog2?
>>
>> Could you please be more specific? Which commands do you mean exactly?
>>
>
> Maybe I'm not understanding the confusion, but I mean the fastlog
> method in evalf:
>
> ###
>>>> from sympy import *
>>>> evalf.fastlog((0, 1, 2, 1)) #i.e. log(2**2)/log(2) = 2
> 2
> ###
>
> Since fastlog is computing the log2 approximation it makes sense to me
> to call it fastlog2 rather than fastlog.
>
> There is no test for this function. I've remedied that and also
> corrected the function to compute the power of an exact power of 2
> properly; previously it was off by 1 in the result. Since the mantissa
> of all mpf numbers provided by mpf is suppose to be odd (according to
> the documentation) then when the number is an exact power of 2, the
> mantissa will be 1, the bitcount of it will also be 1 and the
> approximation (exponent + bitcount) is off by 1: the bitcount should
> not be added in this case.

Ah, yes, I think it should be called fastlog2. Fredrik, what do you think?

>
> If I can figure out how to push my changes to git I will send the
> patch for this by that means. For now, I didn't change the name, I
> only corrected the function and added tests.

Yes, I am on IRC to help you out with github.  Once you get it
working, you'll see that it is easy.

Ondrej

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"sympy" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [email protected]
For more options, visit this group at http://groups.google.com/group/sympy?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to