[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Re: [avr-libc-dev] regression test preview

From: Peeter Vois
Subject: Re: Re: [avr-libc-dev] regression test preview
Date: Tue, 15 Nov 2005 11:01:07 +0200

My opinion is that it is safe to add MUL method into release.

This test shows that MUL method is working precisely with least bit match. It 
does not test the edge effects in purpose. The edge effects (nan, +-inf) are 
another matter for now.

I've seen that when to enlarge the sequential multiplication test from 100 
iterations to 1000 then somewhere AVR will get result nan but PC expects inf.

I've been thinking that we can add this package to new repository into avr-libc 
project. I agree this. If the package will find its home then the versioning 
(something else than 0.0 :) can find place.

It would be helpful when people try to run it and see the problems that can 
raise with the script. It is working on FC4 with updates, but can fail 
somewhere else.

Best Regards ...

Saatja: "Anatoly Sokolov" <address@hidden>
KuupƤev: 14.11.2005 23:47:49
> Hello Peeter.
> > I had some time and I prepared regression test preview package:
> > makes random multiplication test
> > makes sequential multiplication test
> >
> What result of execution of these tests? What your opinion on adding
> the new
> version of MUL function in avr-libc?
> Anatoly.

-- Everyday.com reklaam --  
Tasuta ja kindel e-post: Everyday.com 

reply via email to

[Prev in Thread] Current Thread [Next in Thread]