Re: [eigen] 3.3-alpha1 for mid november?

On Mon, Dec 1, 2014 at 4:16 PM, Christoph Hertzberg <chtz@xxxxxxxxxxxxxxxxxxxxxxxx> wrote:

Apparently, we missed "mid November" but alpha1 would still be possible this year. But we need at least go through all bugs/feature requests blocking 3.3 and decide how to solve them or to postpone them to 3.4.

Right, November has been much more busy than expected.

Regarding input operators we mostly agreed on its behavior. If nobody volunteers to implement the auto-resize part at the moment, we could consider postponing that part (and raise an "not implemented yet" assertion for zero-sized objects at least for alpha1).

Fine to me.

Regarding evaluators, I would appreciate, if we can make the tests compile on <2GB machines again (but no need to block alpha1 for that). Gael, you said memory requirements should decrease after some clean-up?

Did I? If so, I'd not be so affirmative anymore.. We'll see. Am I correct that they currently concern the *SVD unit tests and a few matrix-function unit tests? Perhaps we could try to split them..

For alpha1, we also have to fix the remaining compilation/runtime issues, and it would also be nice to have:

- 877

- 876 (at least a workaround)

Regarding feature requests, I would already defer the following ones to 3.4:

- 751 (numerically robust AMD)

- 707 (inplace decomposition)

- 612 (inplace triangular product/inverse)

- 231 (STL iterators)

- 166 (private allocator)

- 68 (blocked column pivoting QR)

The following twos would be very nice to have, but I'd rather release 3.3 without them and have a 3.4 a very few months later with these two features "only":

- 437 (reshape)

- 329 (indexed views)

Cheers,

Gael

Cheers,
Christoph

On 17.10.2014 18:04, Gael Guennebaud wrote:

Hi,

I'd like to discuss about a release plan for the next 3.3 version.
First, let me recall that the devel branch includes several major new
features:

- AVX/FMA
- D&C SVD (see other thread)
- a novel product kernel
- novel _expression_ template engine

All these features are pretty well advanced without remaining API issues
for normal usage, so I think it's time to plan an alpha1 release asap, for
instance around mid november? This release would be synchronized with a
feature freeze: other features could be added before the alpha1 release, at
the condition they are sufficiently mature and mostly need internal
cleaning/documentation. This probably means that most of the "enhancement"
bug entries attached to 3.3 will be postponed to 3.4.

Then, a beta1 could follow within a month or two. For beta1, all the minor
internal design decisions that are still in suspens regarding the
_expression_ evaluator will have to be resolved.

cheers,
Gael

--
----------------------------------------------
Dipl.-Inf., Dipl.-Math. Christoph Hertzberg
Cartesium 0.049
Universität Bremen
Enrique-Schmidt-Straße 5
28359 Bremen

Tel: +49 (421) 218-64252
----------------------------------------------