Re: [Boost-users] R: R: R: [multi-array] Performance? Difference from GIL or uBLAS?

From: joel falcou (joel.falcou_at_[hidden])
Date: 2008-08-07 15:44:31

> Try to do the pointer access with the simple "res +=
array[ti][tj] ;"
> instruction.
The results from my
earlier post was from this. The res = uni() is not the core of the
computation, just the array initialisation :) The greater the computation,
the smaller the overhead will be IMHO. For a computation like

res + = array[i][j]*cos(array[i][j]) 
versus its while( begin++)
equivalent, the overhead is never greater than 1%

> In a
program where I use much images and arrays the real gain was
>
switching few functions to handwritten assembler code. 6 times faster
than
> an already good c++ algorithm. All the rest of the
application is standard
> elegant and "slow" code. But
it doesn't impact the overall performance.

I perfectly agree
but I think that in most case, going down that low level is not needed,
vene in image processing. i would rather take a few minutes to SIMDify a
code if possible than rewriting it in inline assembly.

Anyway,
I'm onto writing this small multi_array with indexing to see how it fares.