testing code coverage link

[Bob Carpenter]

I saw this link on the Eigen list and thought it'd
be useful for us, though I think we already knew this:

http://stackoverflow.com/questions/9666800/getting-useful-gcov-results-for-header-only-libraries/19792966#19792966

- Bob

[Marcus Brubaker]

On a complete tangent, but this email reminded me of it: tensor support has been merged into Eigen unsupported: http://listengine.tuxfamily.org/lists.tuxfamily.org/eigen/2015/02/msg00019.html

This is relevant for us as it could a) simplify our mix of arrays and matrices into a single type (reducing duplicated code and complexity of supporting operations on both data types) and b) give us a lot more flexibility with indexing with relative ease.  It may also have other uses with things like kronecker products and who knows what else.

Cheers,

Marcus

--
You received this message because you are subscribed to the Google Groups "stan development mailing list" group.
To unsubscribe from this group and stop receiving emails from it, send an email to stan-dev+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Bob Carpenter]

I saw. I wrote an encouraging note to their list
+1-ing the merge, saying we could use it for third-order
derivative tensors instead of the awkward vector<MatrixXd>.

I hadn't thought about getting rid of our std::vector
style arrays, but that's a great idea. I think they bound
the order, but it'll probably be enough for Stan.

It'll be less efficient for slicing than
std::vector (can't return a reference to a row), but tighter
in memory and better at slicing off columsn if we can get
views working.

If we can reduce code duplication, that'd be a win even
for a small performance hit.

- Bob

> To unsubscribe from this group and stop receiving emails from it, send an email to stan-dev+u...@googlegroups.com.

> For more options, visit https://groups.google.com/d/optout.
>

> --
> You received this message because you are subscribed to the Google Groups "stan development mailing list" group.

> To unsubscribe from this group and stop receiving emails from it, send an email to stan-dev+u...@googlegroups.com.

[Ben Goodrich]

On Monday, February 16, 2015 at 8:55:57 PM UTC-5, Bob Carpenter wrote:

I saw.  I wrote an encouraging note to their list
+1-ing the merge, saying we could use it for third-order
derivative tensors instead of the awkward vector<MatrixXd>.

I agree, although my understanding is that there is only limited tensor functionality unless compiled under C++11.

Ben

[Daniel Lee]

[back on topic]

I didn't know about this... it seems like a useful endeavor after we're done with this refactoring.

- Bob