Re: [GSoC] the separate option for all dimensions

[Tobias Grosser <tobias at grosser dot es>]

On 06/08/2014 17:21, Roman Gareev wrote:
> I've tested the modified version of Graphite using the gcc test suite
> and haven't found out new failed tests.
>
> However, pr35356-2.c is still not suitable for testing. The ISL AST
> generated from its source code doesn't contain MIN or MAX.
>
>      if (k <= -1) {
>        for (int c1 = 0; c1 < n; c1 += 1)
>          S_7(c1);
>      } else if (k >= n) {
>        for (int c1 = 0; c1 < n; c1 += 1)
>          S_7(c1);
>      } else {
>        for (int c1 = 0; c1 < k; c1 += 1)
>          S_7(c1);
>        S_6(k);
>        for (int c1 = k + 1; c1 < n; c1 += 1)
>          S_7(c1);
>      }
>
> Should we make pr35356-2.c to be similar to isl-codegen-loop-dumping.c
> by replacing âMIN_EXPR\[^\\n\\r]*;" and "MAX_EXPR\[^\\n\\r]*;" with
> the regexp, which contains the the above-mentioned isl ast?

Checking for this specific AST may cause failures with future versions 
of isl that choose a different schedule. Could you write a regular 
expression that checks that there is no if-condition contained in a for 
loop? I think this best models the issue that was addressed in the 
original bug report.

(Also as this test then becomes isl specific I propose to just force the 
use of isl in the run line).

Cheers,
Tobias