Boost logo

Ublas :

Subject: Re: [ublas] [bindings] Workspaces
From: Thomas Klimpel (Thomas.Klimpel_at_[hidden])
Date: 2008-12-11 04:31:52

Jesse Maning wrote:
> The lack of an overloaded function for specifying an iwork workspace
> was an oversight on my part, there should be an overload for it.

The intended interface is more along the line of what Rutger described:
 * minimal
 * optimal
 * user-defined
so there should be no separate overload for iwork. Instead, in case of
user-defined, also iwork must be given be the user.

> When I was implementing the bindings i remember being a little
> about the iwork and i didn't see the heevd implementation so i settled
> on creating it with the size specified in the intel MKL reference
> using the function ilaenv with a parameter of 9 for getting the number
of subproblems.

The heevd implementation didn't exist when you wrote your bindings, so
there was no fault from your side. I only had to make it clear that I
was well aware that gelsd didn't use the intended workspace interface,
even so I added it only recently.

> For the iwork parameter should the user be able to select a
minimum/optimal configuration like the other workspaces?

No, it is always of fixed size.