Done in
https://github.com/grame-cncm/faust/commit/8fb5a2b47adb6a7efa125ee5be224b08ac371ec8
Stéphane
> Le 28 août 2023 à 20:27, Oleg Nesterov <o...@redhat.com> a écrit :
>
> On 08/27, Julius Smith wrote:
>>
>> Ok great - sorry I didn't read the whole thread carefully - just a drive-by
>> observation
>
> Nevertheless I was wrong, thanks for correcting me.
>
> And I guess I confused Stephane, sorry. The change in atan2prim.hh from
> faust commit 26537a42f0fe55 ("Optimize atan2(x,x) as PI/4. Set version to
> 2.59.3") must be reverted.
>
> Oleg.
>
>> On Sun, Aug 27, 2023 at 11:08 AM Oleg Nesterov <o...@redhat.com> wrote:
>>
>>> On 08/27, Julius Smith wrote:
>>>>
>>>>> note that atan2(fTemp0, fTemp0) == %pi/4
>>>>
>>>> That's only true if fTemp0 is positive.
>>>> For negative fTemp0, it's 5 * %pi / 4, or - 3 * %pi / 4, depending on
>>> your
>>>> choice of domain.
>>>
>>> Yes, and if fTemp0 == 0 then atan2() is undefined.
>>>
>>> And this means that the recent faust commit 26537a42f0fe55c7d471c is wrong,
>>>
>>> process = _ <: atan2;
>>>
>>> outputs
>>>
>>> output0[i0] = FAUSTFLOAT(0.7853982f);
>>>
>>> Yes, sorry for confusion, I just tried to explain how it can be used in the
>>> pure faust code, and didn't even think about correctness.
>>>
>>> But let me repeat:
>>>
>>>> I didn't even think about optimizing sigBinOp's. I used atan2
>>>> only for illustration and as a trivial test-case.
>>>
>>> Oleg.
>>>
>>>
>>
>> --
>> "Anything that can be automated should optionally be"
>
_______________________________________________
Faudiostream-devel mailing list
Faudiostream-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/faudiostream-devel
