On 10/19, Stéphane Letz wrote:
>
> So we test the backend behaviour, not the generated code by itself.
OK... and I guess we can't rely on -norm1 to verify that the generated
code will be "the same". It's a pity ;)
Thanks!
Oleg.
___
Faudiostream-deve
> Le 19 oct. 2023 à 21:39, Oleg Nesterov a écrit :
>
> On 10/19, Stéphane Letz wrote:
>>
>> Fixed in just pushed 2.68.5
>
> looks like the commit 637b63bfdfa1f6b33e38 does the trick, thanks! ;)
>
> but... could you help please? I have never looked into the tests/
> directory and I am puzzled
On 10/19, Stéphane Letz wrote:
>
> Fixed in just pushed 2.68.5
looks like the commit 637b63bfdfa1f6b33e38 does the trick, thanks! ;)
but... could you help please? I have never looked into the tests/
directory and I am puzzled.
This commit adds the new tests/impulse-tests/dsp/math_simp.dsp file,
Fixed in just pushed 2.68.5
Stephane
> Le 19 oct. 2023 à 19:04, Oleg Nesterov a écrit :
>
> On 10/13, Stéphane Letz wrote:
>>
>> Yes, Yann is looking at some misoptimisations that should fixed.
>
> Thanks, I see the new commits... the unnecessary *(-1) has gone.
>
> But just FYI, this trivia
On 10/13, Stéphane Letz wrote:
>
> Yes, Yann is looking at some misoptimisations that should fixed.
Thanks, I see the new commits... the unnecessary *(-1) has gone.
But just FYI, this trivial example
import("stdfaust.lib");
k = ma.SR;
process(x) = -k, -k*x;
still compi
Yes, Yann is looking at some misoptimisations that should fixed.
Stéphane
> Le 13 oct. 2023 à 21:22, Oleg Nesterov a écrit :
>
> On 10/13, Oleg Nesterov wrote:
>>
>> OK. to be honest I have some concerns about this patch... In particular I
>> still think
>> that the special "ksub && izZero"
On 10/13, Oleg Nesterov wrote:
>
> OK. to be honest I have some concerns about this patch... In particular I
> still think
> that the special "ksub && izZero" case in simplification() should die, but
> perhaps
> I am wrong. I'll recheck later.
If nothing else...
process(x) = (0 - (0 - x
