|Re: [eigen] Sparse -- Diagonal product|
[ Thread Index |
| More lists.tuxfamily.org/eigen Archives
- To: eigen@xxxxxxxxxxxxxxxxxxx
- Subject: Re: [eigen] Sparse -- Diagonal product
- From: Benoit Jacob <jacob.benoit.1@xxxxxxxxx>
- Date: Wed, 20 May 2009 14:20:31 +0200
- 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=kHuy3LasLablhWB1NCNsFpBMfKmhhLZSvz65fb7IV4U=; b=L0XhUeUHXMfeS3eCbxtAh8rwwXY7bqqNpimQYunQTGeLGWqWr5Njte0SfKTriNu8I1 3AT6HHnIU4Iobt/0gLPrfBp3WEZ+wZ92V7oSJv+d411jOtqTS7J1RFKFLTZC3MmJqgot 059rIKujB6hxI1HX8VDbzZ5KMYTdnc2aEbb+g=
- 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=tUcf5Q0fE8ZtPtoRmE6Owh3kPVDkMfjk/NmKkHXWBsJgJA2Mu8X6k568DMcfumdSa5 ZnsL2AWC5chsNVnTw3E12C6gg1oCb5egZskWRk9OmTz085QTt0Ljd+umxpk7D+He+PiX Y1IIDD190d6DwwXLUxri5EQT7jm8uQMvj8Z+M=
Aaahh forget _all_ of what I said!
I misread your email -- thought that you were talking about diagonal
matrices evaluating to dense matrices in Core/.
Only take into account Gael's answer :)
2009/5/20 Benoit Jacob <jacob.benoit.1@xxxxxxxxx>:
> Actually I am just working on this at the moment (see recent threads
> on diagonal matrices).
> As Gael said this is changing.
> Normally I'd happily let you handle this but this one is really
> tightly integrated with several other core issues.
> 2009/5/20 Justin Carlson <justinca@xxxxxxxxxx>:
>> I was a bit surprised today to figure out that Eigen current evaluates
>> the product of a diagonal matrix with a sparse matrix as a dense
>> Is this intentional? Changing the result to be sparse (which seems
>> sane to me, particularly given my current usage of the library) seems
>> like the right thing to do, but it would represent an API change.
>> Is there a design rationale for this, or is it just that no
>> appropriate specializations have been implemented yet for this case?
>> If there aren't objections to changing the behavior, I'll see about
>> submitting a patch.