
How to perform a element wise product of two vecotr (or matrix)
Rossi88 Feb 9, 2010 6:16 PM (in response to Rossi88)In the MKL (Math Kernel Library provided by Intel) the function that I search have name like
vdsum
vdmul
vdsub
and so on, that respectively made the sum, the multiplication, the subtraction ... element by element of two vector of double element.
Tthere isn't nothing of similar in ACML?

How to perform a element wise product of two vecotr (or matrix)
chipf Feb 9, 2010 8:11 PM (in response to Rossi88)ACML does not provide these functions.
AMD provided a library, now called FrameWave, that is available on source forge.
The link to it can be found on our developer central web page 
http://developer.amd.com/cpu/Libraries/framewave/Pages/default.aspx
There is a link to the online documentation. There are a set of vector arithmetic and logic functions in the image processing library section.

How to perform a element wise product of two vecotr (or matrix)
Rossi88 Feb 20, 2010 6:33 AM (in response to chipf)Thanks a lot for the answer.
can you say to me if this framework allow to increase perfomance? or make only easier the operation of multiplication, subtraction and so on element by element with a vector?
in your opinion you think that this framework would be useful for me?

How to perform a element wise product of two vecotr (or matrix)
chipf Feb 24, 2010 1:44 PM (in response to Rossi88)Can't answer for certain. The Framewave project does include performance optimized implementations.
It will most likely be worth the effort to download and try it out.


