Age | Commit message (Collapse) | Author | Files | Lines |
|
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
The result computed by isl_hash_dim may affect the order of maps
in a union map, which may in turn affect the result of AST generation.
In particular, we would generate slightly different output depending
on endianness, causing the tests to fail on big endian machines since
the generated output is compared to output generated on a little endian
machine.
Instead of hashing in the dimensions of the space as integers,
we now only hash in the least significant byte.
This has some effect on the results of AST generation, but the output
should now be the same on little endian and big endian.
Reported-by: Richard Biener <rguenther@suse.de>
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: 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>
|
|
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
We cannot simply export the private isl_space_is_domain however, since
that one is also used in cases where the first argument is a parameter
space and the second argument is a set space.
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
The original code would only take into account the number of dimensions
and the tuple name.
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: Sven Verdoolaege <skimo@kotnet.org>
|
|
In particular, if we are inserting parameters, then we should also
insert them in any nested spaces.
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: Sven Verdoolaege <skimo@kotnet.org>
|
|
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
A nested space is supposed to have the same parameters as
the nesting space, so if the id of a parameter in the nesting
space space changes, it has to be changed in the nested spaces too.
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: Sven Verdoolaege <skimo@kotnet.org>
|
|
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
Before, no distinction was made internally between spaces of maps,
sets and parameter spaces. In particular, an isl_space with zero
input and output dimensions could have been any of those.
When printing isl_sets and isl_maps, we could keep track of whether
the original object as a set or a map, but parametric domains and
zero-dimensional domains were still printed identically.
Now we explicitly keep track of whether a space is supposed to be
that of a map, set or parametric set. This means for example that
"[N] -> { [] : N >= 0 }" and "[N] -> { : N >= 0 }" are no longer considered
to be identical.
This may break some code that was using one space in place of another,
especially since until recently there was no official way of creating
a parameter space.
The main advantage is that we can now recognize parameter spaces
and treat them appropriately. In particular, the domains of affine
expressions can now be correctly identified as being either (possibly
zero-dimensional) sets or parameter domains.
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
A space is obviously equal to itself.
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
Before, we would only accept parameter spaces.
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|
|
The old name was confusing because the name suggested that the object
represents a single dimension, while in fact it represents an entire space.
The documented isl_dim based names are for backward compatibility.
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>
|