| Re: [eigen] sse transcendental functions |
[ Thread Index |
Date Index
| More lists.tuxfamily.org/eigen Archives
]
- To: eigen@xxxxxxxxxxxxxxxxxxx
- Subject: Re: [eigen] sse transcendental functions
- From: Rohit Garg <rpg.314@xxxxxxxxx>
- Date: Wed, 25 Mar 2009 22:40:26 +0530
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:received:in-reply-to:references :date:message-id:subject:from:to:content-type :content-transfer-encoding; bh=4kzJMCv0xS0hqFihw3YBOJr5KnKuxrs23FZX/D4bCqg=; b=eqFh7SvMvV28tDAfZ/alQI6GT2H8taISzPlpvk428P20DtSoQ5wJ9Ah62nQI573epY z2Gy5q0/3GPcK9bYTd3F9FScZbtBdO9YQM98p7u8wOvyGta1386CWSuHxdRYVGt/djs4 kZalLjdGKfTJhYD+7JQ1/mx9ER1eGT2IO0WWQ=
- Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :content-type:content-transfer-encoding; b=FCXj42ikpEaGxGJk2cm7X0+sXjZraAtCMd31as/yYQp12uepRDbAk83guhU9iHn9UR 7EMETfo9QVezDquitoxNgmH8l7KK/gCP74nDSNxmKddmkwaiy4ZhX1ofkTw/jibzDx2l 2CJPvYWo6epE+U62UT6jMvZxDuOBoqlIgTjNE=
I think they do convert the values into this range and then calculate
the values. Any way their sincos4f would be useful.
On Wed, Mar 25, 2009 at 10:37 PM, Gael Guennebaud
<gael.guennebaud@xxxxxxxxx> wrote:
> ok, so this actually corresponds to the SIMD lib by Sony, and their
> SSE sinf4 is indeed slightly faster (200M calls versus 177M) but it is
> only valid in the range [-pi/4 : pi/4] so it is quite useless for us.
> Note that both implementations are based on exactly the same principle
> and the speed difference is 100% explained by the handling of a much
> wider range of values
>
> gael.
>
> On Wed, Mar 25, 2009 at 5:36 PM, Rohit Garg <rpg.314@xxxxxxxxx> wrote:
>> I'll dig them up
>>
>>
>> On Wed, Mar 25, 2009 at 10:06 PM, Gael Guennebaud
>> <gael.guennebaud@xxxxxxxxx> wrote:
>>> where are they ? I can only find btsin and btCos which simply call the
>>> standard C functions....
>>>
>>> On Wed, Mar 25, 2009 at 3:35 PM, Rohit Garg <rpg.314@xxxxxxxxx> wrote:
>>>> please consider benching these against the sin, cos, acos, sincos
>>>> implementations in bullet physics sdk. It is bsd licensed and may be
>>>> faster.
>>>>
>>>> --
>>>> Rohit Garg
>>>>
>>>> http://rpg-314.blogspot.com/
>>>>
>>>> Senior Undergraduate
>>>> Department of Physics
>>>> Indian Institute of Technology
>>>> Bombay
>>>>
>>>>
>>>>
>>>
>>>
>>>
>>
>>
>>
>> --
>> Rohit Garg
>>
>> http://rpg-314.blogspot.com/
>>
>> Senior Undergraduate
>> Department of Physics
>> Indian Institute of Technology
>> Bombay
>>
>>
>>
>
>
>
--
Rohit Garg
http://rpg-314.blogspot.com/
Senior Undergraduate
Department of Physics
Indian Institute of Technology
Bombay