| Re: [eigen] Malloc-free dynamic matrices |
[ Thread Index |
Date Index
| More lists.tuxfamily.org/eigen Archives
]
- To: eigen@xxxxxxxxxxxxxxxxxxx
- Subject: Re: [eigen] Malloc-free dynamic matrices
- From: Benoit Jacob <jacob.benoit.1@xxxxxxxxx>
- Date: Thu, 4 Mar 2010 17:02:37 -0500
- 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; bh=tNJk9Az/GJaP+Kd5yMCuKerL82d3ocTtNzMiAiUgZWQ=; b=BvQ+SxoDntDwh/G3bb8I/6ILOeJVOt3qAPqIszIRbX7nNmLUINc1zoh+ky9gRlGQBB TyUsW3ZMTQknv/jfqffFUt9mGOOayu/aOK2IM8obKCXuVuCN8YthLOAcsLgBKWbVAbJT 5U5YNUwJTCm3nYYhUtyb57YvJX19HtWtI6urQ=
- 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; b=m9zqrQaMajICNECO1YCAp8+4H0BFMNeEkF29CeUGvsRQs498QcBHp+48VGd3sRWCKA iydE9xhKPdgR7KjdyeB3S2x+7Oo4sQnh+NV2HYTpeSFG+dEcQ6VssvlgHcVMwWwU3Pu7 M5mwzI60PXBARA0IG+hwct8PEXQfRrGe95pmg=
2010/3/4 Gael Guennebaud <gael.guennebaud@xxxxxxxxx>:
> on the other hand, if manage_cache_size is a template (e.g., to have one per
> scalar type) then that does not work across multiple object files.
If this is really a problem then this is an argument in favor of
offering a bytes-based API for setting cache sizes, and computing the
actual matrix block sizes at runtime.
If the main argument against that was the cost of calling sqrt() then
here's a quick remark: for integers, a rough estimate of sqrt is very
quick to compute, once you know the number of bits N you just take x
>> (N/2)... something based on that might be accurate enough and very
fast... but it's not even clear to me at all that the cost of sqrt is
a problem.
Benoit
>
> gael
>
>>
>> gael.
>>
>>>
>>> gael.
>>>
>>>>
>>>> Benoit
>>>>
>>>>
>>>
>>
>
>