Ok, thank you very much.
Jacob (Jack) Gryn schrieb:
> Done, and committed to CVS.
>
> I tested the gmd one, I'm assuming the others will work, the only difference
> is the complex one where it did .r and .i separately (but in the same loop).
I agree, the complex case should be fine as well.
By the way, did you think about the work size variable of your
LaLUInverseIP code?
http://sourceforge.net/mailarchive/message.php?msg_id=11404933
Christian
>
> Jack
>
>
>>-----Original Message-----
>>From: lap...@li... [mailto:lapackpp-devel-
>>ad...@li...] On Behalf Of Christian Stimming
>>Sent: April 18, 2005 4:03 PM
>>To: Jacob (Jack) Gryn; lap...@li...
>>Subject: Re: [Lapackpp-devel] Operator+=
>>
>>Am Montag, 18. April 2005 19:45 schrieb Jacob (Jack) Gryn:
>>
>>>Just FYI, I was referring to operator+=(double) or operator+=(complex),
>>>etc. (not matrix addition which you can do with Blas_Add_Mult).
>>
>>These are fine and you can add them as you wish. They are quite clear
>>about
>>what will happen and which matrices/vectors get modified. (I'm actually
>>wondering why I never needed those... whatever. Simply add them.)
>>
>>Regards,
>>
>>Christian
|