← Back to team overview

yade-dev team mailing list archive

Re: float, double, the DOUBLE_PRECISION flag

 

I could do the tests but I have still few things to code before I can 
run the simulations I want.
Is it a problem to just remove the float/double switch mechanism and fix 
Real to double.
That way, we wouldn't have to change our code when Real appears 
somewhere, and it would let us an easy way to test something like double 
vs longdouble in the future.
I forsee that the answer to the first question will be "yes"...  Vaclav 
will have to explain that to me again... ;-)

Bruno


Vaclav SMILAUER a écrit :

>Oh, thanks...
>
>I will wait with that until wm3 is cleaned up and a few weeks pass 
>without major regressions. Then we can go ahead.
>
>It could be interesting to compare double vs. long double, although it 
>could not be done easily. Perhaps before Real is removed, typedef it as 
>long double... Is it worh the effort? Is wm3 long-double-aware?
>Is someone interested in running two simulations and compare the 
>results, as it was done for floats vs. doubles (I have no time for that, 
>sorry)? If not, I will go ahead.
>
>Vaclav
>
>  
>
>>Hi Vaclav
>>
>>I've been thinking whole yesterday, and finally I must give your
>>honor back and admit that your arguments were correct. I hope that
>>you will survive with so stubborn leader ;)
>>
>>Go ahead and remove Real.
>>
>>    
>>
>
>_______________________________________________
>yade-dev mailing list
>yade-dev@xxxxxxxxxxxxxxxx
>https://lists.berlios.de/mailman/listinfo/yade-dev
>
>
>
>  
>


-- 
_______________
Chareyre Bruno
Maître de conférence

Institut National Polytechnique de Grenoble
Laboratoire 3S (Soils Solids Structures) - bureau I08
BP 53 - 38041, Grenoble cedex 9 - France     
Tél : 04.56.52.86.21
________________

_______________________________________________
yade-dev mailing list
yade-dev@xxxxxxxxxxxxxxxx
https://lists.berlios.de/mailman/listinfo/yade-dev



References