Test Case Failed

[bondhugula]

Test Case Failed

[bondhugula]

Which version of Pluto are you using? This test case is meant/known to fail due to an aspect of the Pluto's lazy version that hasn't been implemented -- this is a contrived and a very rare case. Also, in such a case, Pluto will default to the identity transformation and go ahead with other steps. With the git version, you shouldn't see a failure. This should be the output

$ ./polycc test/multi-stmt-lazy-lin-ind.c
[pluto] compute_deps (isl)
[pluto] Number of statements: 2
[pluto] Total number of loops: 4
[pluto] Number of deps: 3
[pluto] Maximum domain dimensionality: 2
[pluto] Number of parameters: 2
[pluto] WARNING: working with original (identity) transformation (if they exist)
[pluto] Affine transformations

T(S1): (0, i, 0, j, 0)
loop types (loop, loop, loop, loop, loop)

T(S2): (0, i, 0, j, 1)
loop types (loop, loop, loop, loop, loop)

[pluto] using Cloog -f/-l options: 1 5
[CLooG] INFO: 1 domains have been blocked.
[CLooG] INFO: 3 dimensions (over 5) are scalar.
[pluto] Output written to multi-stmt-lazy-lin-ind.pluto.c

[pluto] Timing statistics
[pluto] SCoP extraction + dependence analysis time: 0.015447s
[pluto] Auto-transformation time: 0.005465s (constraint solving: 0.000000s)
[pluto] Code generation time: 0.002322s
[pluto] Other/Misc time: 0.005024s
[pluto] Total time: 0.028258s
[pluto] All times: 0.015447 0.005465 0.002322 0.005024

[skewer]

Thank you for your kind reply!
I was trying to compile master branch of polymage, so I used the commit id [145fdf4].
The test will pass if no argument is provided to polycc, but it failed with argument "--partlbtile --cloogsh --nocloogbacktrack --debug", when I use "make test " to test the build.
I also tried the same argument with the git version, and it failed too. Does it mean this test is meant to fail with a successful build? Maybe It is meaningless to use this argument in this case.

[bondhugula]

Yes, that test is meant to fail. Please ignore it.

[skewer]

Thanks!