Age | Commit message (Collapse) | Author | Files | Lines |
|
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
Signed-off-by: Tobias Grosser <tobias@grosser.es>
Signed-off-by: Andreas Kloeckner <kloeckner@cims.nyu.edu>
Signed-off-by: Todor Stefanov <stefanov@liacs.nl>
Signed-off-by: Sven van Haastregt <svhaastr@liacs.nl>
Signed-off-by: Isabelle Ryl <isabelle.ryl@inria.fr>
Signed-off-by: Mythri Alle <mythri.allel@gmail.com>
Signed-off-by: Wim De Clercq <Wim.DeClercq@lrd.kuleuven.be>
Signed-off-by: Anne Cormier <Anne.Cormier@ens.fr>
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
This commit fixes two closely related problems.
Due to a typo, isl_set_dim_residue_class would only consider one of
the basic sets in the input set. The information obtained from the
individual basic sets was also combined incorrectly.
In particular, the code would insist that all basic sets have the
same residue if they have the same modulo and not consider adjusting
the modulo of the combined basic sets to accommodate for the difference
in residues.
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
Although the *_fast_* functions are certainly meant to be faster than
their unqualified alternatives, they are faster only because they are
not complete. The "plain" qualification is hopefully better at conveying
the idea that these functions only consider the obvious cases.
We keep some of the *_fast_* names for backward compatibility.
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
We need to turn on the nostdinc option of automake as otherwise
it would include -I$(top_builddir)/include/isl in DEFAULT_INCLUDES
because of
AC_CONFIG_HEADERS(include/isl/config.h)
in configure.ac.
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Before, when no residue class could be detected, *modulo was sometimes
set to 0 and sometimes to 1. Now *modulo is always set to 1 in these
cases, except when the specified dimension has a fixed constant value,
in which case *modulo is set to 0 and *residue to the fixed value.
|
|
|
|
|
|
|
|
|
|
|
|
The old version would sometimes remove valid solutions.
Since isl_mat_parameter_compression is used during simplification
on practically any set containing existentially quantified variables
involved in equalities, any such set could get corrupted.
|
|
|
|
|
|
|
|
All of isl_basic_set, isl_set, isl_basic_map and isl_map stored
information about the dimension directly. This information is
now shared in a common isl_dim structure.
Direct accesses to the removed fields have been replaced by calls
to getters.
|
|
|
|
|
|
Since isl_sets and isl_maps now know which isl_ctx they belong
to, the isl_ctx argument of many functions can be dropped.
|
|
These changes are needed if we want isl_basic_set_sample
to always return an actual point. Without them, the lexmin
may be unbounded and we effectively get a ray.
|
|
|