Rutger ter Borg wrote:
IMHO, for a next generation C++ linear algebra library to be really successful, I think the following properties are desired
* split containers from expressions
That's a given no ?
* split algorithms from expressions, i.e., enable pluggable/customisable back-ends into the expression templates
Ultimately, the idea could be to have a way to fall back to BLAS or w/e exists ona given platform when specified.
* achieve performance of FORTRAN-based programs (dominant in HPC)
By handling low level parallelism construct, we can even out perform it. Add also multi-parallel architecture support. Next 10 years will eb full of dense integrated multicores everywhere. So better take time to be prepared for that. -- ___________________________________________ Joel Falcou - Assistant Professor PARALL Team - LRI - Universite Paris Sud XI Tel : (+33)1 69 15 66 35