Re: [ublas] Performance with SSE2

From: Gunter Winkler (guwi17_at_[hidden])
Date: 2007-03-12 11:25:32

On Monday 12 March 2007 16:16, Preben Holm wrote:
> Has there been made any other flags by boost which could optimize
> performance by using SSE2 vector extensions!

AFAIR you only gain performance benefits if the compiler is able to 
automatically vectorize computations. Unfortunatly are the expression 
template constructions quite complicated. So most compilers do not recognize 
possible optimizations. Second, there were (some years ago) some explicit 
loop unrolling tests. Look for BOOST_UBLAS_USE_DUFF_DEVICE inside the code. 
(Tell me if you manage successful compilation with 
BOOST_UBLAS_USE_DUFF_DEVICE defined.)

I think, if you look for maximal performance, you have to use the 
BLAS-bindings inside some critical routines. All other access patters are to 
complicated to benefit from SIMD anyway.

mfg
Gunter