On 05/10/10 16:31, Don wrote:
Gareth Charnock wrote:
On 05/10/10 14:41, Michael Chen wrote:
I remember that one of D's goal is easy scientific computation.
However I haven't seen any linear algebra package for D2. My work
heavily relays on all kinds of matrix stuff (matrix multiplication,
factorization, linear system etc). I like D and am willing to work
with D. However without these facilities I can hardly start. I'd like
to have a matrix library of which the API is kind of like Matlab.
Is there anybody working on this or planning to work on this?

Regards,
Michael

I've attempted this in the past, however, each time I manage to work
up renewed enthusiasm I keep running into compiler bugs that put me
off. My current feeling is I should sit back at wait for the language
to mature a little.

However, once D does mature there are some neat tricks that would be
possible in a matrix library. For a start you could easily build
something like ATLAS with all the tuning parameters being passed as
template parameters. Another neat trick might be verifying that a
matrix is othoginal/unitary/whatever with invariants. Again a static
if and a template parameter makes this feature really trivial to add.

As for things that aren't vaporware,
http://www.dsource.org/projects/mathextra/browser/trunk/blade might be
a good place to start.

Please don't. That was a proof of concept, which had a big influence on
D's array operations and the design of D's metaprogramming support. All
of the ideas from BLADE will eventually be completely incorporated into
array operations.
Basically, right now D has DAXPY and SAXPY from BLAS1, implemented as
array operations. But it doesn't have anything else at present.
Okay, sorry, I may have somewhat misunderstood the purpose of that library, I hadn't really looked at it in great detail, I just knew it was a math library.

Reply via email to