Tobias Grosser 6217e18a7d Add preliminary implementation for GPGPU code generation. ... Translate the selected parallel loop body into a ptx string and run it with the cuda driver API. We limit this preliminary implementation to target the following special test cases: - Support only 2-dimensional parallel loops with or without only one innermost non-parallel loop. - Support write memory access to only one array in a SCoP. The patch was committed with smaller changes to the build system: There is now a flag to enable gpu code generation explictly. This was required as we need the llvm.codegen() patch applied on the llvm sources, to compile this feature correctly. Also, enabling gpu code generation does not require cuda. This requirement was removed to allow 'make polly-test' runs, even without an installed cuda runtime. Contributed by: Yabin Hu <yabin.hwu@gmail.com> llvm-svn: 161239		2012-08-03 12:50:07 +00:00
..
GPGPU	Add preliminary implementation for GPGPU code generation.	2012-08-03 12:50:07 +00:00
MemAccess	CodeGen: Recreate old ivs with the original type	2012-04-03 12:24:32 +00:00
OpenMP	CodeGen: Allow function parameters to be rewritten in getNewValue()	2012-04-01 16:49:45 +00:00
20100617.ll	test: Do not pipe the .ll file into opt	2011-07-05 19:13:21 +00:00
20100622.ll	test: Do not pipe the .ll file into opt	2011-07-05 19:13:21 +00:00
20100707.ll	test: Do not pipe the .ll file into opt	2011-07-05 19:13:21 +00:00
20100707_2.ll	test: Do not pipe the .ll file into opt	2011-07-05 19:13:21 +00:00
20100708.ll	test: Do not pipe the .ll file into opt	2011-07-05 19:13:21 +00:00
20100708_2.ll	test: Do not pipe the .ll file into opt	2011-07-05 19:13:21 +00:00
20100713.ll	test: Do not pipe the .ll file into opt	2011-07-05 19:13:21 +00:00
20100713_2.ll	test: Do not pipe the .ll file into opt	2011-07-05 19:13:21 +00:00
20100717.ll	test: Do not pipe the .ll file into opt	2011-07-05 19:13:21 +00:00
20100718-DomInfo-2.ll	test: Do not pipe the .ll file into opt	2011-07-05 19:13:21 +00:00
20100718-DomInfo.ll	test: Do not pipe the .ll file into opt	2011-07-05 19:13:21 +00:00
20100720-MultipleConditions.c	…
20100720-MultipleConditions.ll	test: Switch to new atomic instructions	2011-11-29 14:51:05 +00:00
20100809-IndependentBlock.ll	…
20100811-ScalarDependencyBetweenBrAndCnd.ll	…
20101030-Overflow.ll	…
20101103-Overflow3.ll	…
20101103-signmissmatch.ll	…
20110226-Ignore-Dead-Code.ll	…
20110226-PHI-Node-removed.ll	…
20110312-Fail-without-basicaa.ll	…
20120316-InvalidCast.ll	Don't allow pointer types in affine expressions	2012-03-16 16:36:47 +00:00
20120403-RHS-type-mismatch.ll	CodeGen: Recreate old ivs with the original type	2012-04-03 12:24:32 +00:00
PHIInExit.ll	test: Do not pipe the .ll file into opt	2011-07-05 19:13:21 +00:00
constant_condition.c	…
constant_condition.ll	test: Switch to new atomic instructions	2011-11-29 14:51:05 +00:00
do_pluto_matmult.c	…
do_pluto_matmult.ll	Don't fail the lli testcases on 32bit platform.	2012-03-26 15:16:48 +00:00
do_pluto_matmult.ll.result	…
do_pluto_matmult___%do.body---%do.end45.jscop	…
do_pluto_matmult___%do.body---%do.end45.jscop.invalid_reverse	…
do_pluto_matmult___%do.body---%do.end45.jscop.valid_reverse	…
loop_with_condition.c	…
loop_with_condition.ll	Don't fail the lli testcases on 32bit platform.	2012-03-26 15:16:48 +00:00
loop_with_condition_2.ll	Don't fail the lli testcases on 32bit platform.	2012-03-26 15:16:48 +00:00
loop_with_condition_ineq.c	…
loop_with_condition_ineq.ll	Don't fail the lli testcases on 32bit platform.	2012-03-26 15:16:48 +00:00
loop_with_condition_nested.c	…
loop_with_condition_nested.ll	Don't fail the lli testcases on 32bit platform.	2012-03-26 15:16:48 +00:00
matmul_vec.c	…
matmul_vec.ll	Regression tests: Adapt the vectorize option change.	2012-05-06 10:22:43 +00:00
matmul_vec___%1---%17.jscop	…
pluto_matmult.c	…
pluto_matmult.ll	tests: Replace . by %s	2012-02-13 12:29:43 +00:00
pluto_matmult___%for.cond---%for.end47.jscop	…
reduction.c	…
reduction.ll	test: Switch to new atomic instructions	2011-11-29 14:51:05 +00:00
reduction_2.c	…
reduction_2.ll	Don't fail the lli testcases on 32bit platform.	2012-03-26 15:16:48 +00:00
scev.ll	test: Do not pipe the .ll file into opt	2011-07-05 19:13:21 +00:00
sequential_loops.c	…
sequential_loops.ll	Don't fail the lli testcases on 32bit platform.	2012-03-26 15:16:48 +00:00
simple_nonaffine_loop.c	Support non-affine access functions in Polly.	2011-12-20 10:43:14 +00:00
simple_nonaffine_loop.ll	Support non-affine access functions in Polly.	2011-12-20 10:43:14 +00:00
simple_vec_assign_scalar.c	…
simple_vec_assign_scalar.ll	Regression tests: Adapt the vectorize option change.	2012-05-06 10:22:43 +00:00
simple_vec_assign_scalar_2.c	…
simple_vec_assign_scalar_2.ll	Regression tests: Adapt the vectorize option change.	2012-05-06 10:22:43 +00:00
simple_vec_call.ll	Regression tests: Adapt the vectorize option change.	2012-05-06 10:22:43 +00:00
simple_vec_call_2.ll	Regression tests: Adapt the vectorize option change.	2012-05-06 10:22:43 +00:00
simple_vec_cast.ll	Regression tests: Adapt the vectorize option change.	2012-05-06 10:22:43 +00:00
simple_vec_const.c	…
simple_vec_const.ll	Regression tests: Adapt the vectorize option change.	2012-05-06 10:22:43 +00:00
simple_vec_impossible.ll	Regression tests: Adapt the vectorize option change.	2012-05-06 10:22:43 +00:00
simple_vec_large_width.ll	Regression tests: Adapt the vectorize option change.	2012-05-06 10:22:43 +00:00
simple_vec_stride_one.c	…
simple_vec_stride_one.ll	Regression tests: Adapt the vectorize option change.	2012-05-06 10:22:43 +00:00
simple_vec_stride_one___%1---%5.jscop	…
simple_vec_stride_x.c	…
simple_vec_stride_x.ll	Regression tests: Adapt the vectorize option change.	2012-05-06 10:22:43 +00:00
simple_vec_two_stmts.ll	Regression tests: Adapt the vectorize option change.	2012-05-06 10:22:43 +00:00
single_do_loop_int_max_iterations.c	test: Switch to new atomic instructions	2011-11-29 14:51:05 +00:00
single_do_loop_int_max_iterations.ll	Don't fail the lli testcases on 32bit platform.	2012-03-26 15:16:48 +00:00
single_do_loop_int_max_iterations___%do.body---%do.end.jscop	…
single_do_loop_int_param_iterations.c	test: Switch to new atomic instructions	2011-11-29 14:51:05 +00:00
single_do_loop_int_param_iterations.ll	test: Switch to new atomic instructions	2011-11-29 14:51:05 +00:00
single_do_loop_ll_max_iterations.c	test: Switch to new atomic instructions	2011-11-29 14:51:05 +00:00
single_do_loop_ll_max_iterations.ll	test: Switch to new atomic instructions	2011-11-29 14:51:05 +00:00
single_do_loop_one_iteration.c	…
single_do_loop_one_iteration.ll	test: Switch to new atomic instructions	2011-11-29 14:51:05 +00:00
single_do_loop_scev_replace.c	test: Switch to new atomic instructions	2011-11-29 14:51:05 +00:00
single_do_loop_scev_replace.ll	Don't fail the lli testcases on 32bit platform.	2012-03-26 15:16:48 +00:00
single_loop.c	…
single_loop.ll	Don't fail the lli testcases on 32bit platform.	2012-03-26 15:16:48 +00:00
single_loop_int_max_iterations.c	…
single_loop_int_max_iterations.ll	test: Switch to new atomic instructions	2011-11-29 14:51:05 +00:00
single_loop_ll_max_iterations.c	…
single_loop_ll_max_iterations.ll	test: Switch to new atomic instructions	2011-11-29 14:51:05 +00:00
single_loop_one_iteration.c	…
single_loop_one_iteration.ll	test: Switch to new atomic instructions	2011-11-29 14:51:05 +00:00
single_loop_param.ll	Don't fail the lli testcases on 32bit platform.	2012-03-26 15:16:48 +00:00
single_loop_uint_max_iterations.c	…
single_loop_uint_max_iterations.ll	test: Switch to new atomic instructions	2011-11-29 14:51:05 +00:00
single_loop_ull_max_iterations.c	…
single_loop_ull_max_iterations.ll	test: Switch to new atomic instructions	2011-11-29 14:51:05 +00:00
single_loop_zero_iterations.c	…
single_loop_zero_iterations.ll	test: Switch to new atomic instructions	2011-11-29 14:51:05 +00:00
split_edges.ll	Codegen: Give split and merge basic blocks better names	2012-02-12 12:09:37 +00:00
split_edges_2.ll	CodeGen: Always name merge block	2012-02-12 12:09:46 +00:00
test.c	…
test.ll	test: Do not pipe the .ll file into opt	2011-07-05 19:13:21 +00:00