Commit Graph

13 Commits

Author SHA1 Message Date
Tobias Grosser 24bb46bc66 [autoconf/cmake] Make sure we detect the latest version of isl.
llvm-svn: 185429
2013-07-02 14:11:32 +00:00
Sebastian Pop f2379e0848 do not require cloog from configure
llvm-svn: 168628
2012-11-26 23:03:41 +00:00
Sebastian Pop 654b7754bc autoconf: isl depends on gmp include files
When libgmp header files are not installed in the default /usr/include location,
configure used to fail to find the gmp headers when testing for isl/ast.h.  This
patch adds the gmp include path to the compiler flags before testing for ISL.

llvm-svn: 168090
2012-11-15 21:20:22 +00:00
Tobias Grosser 0c55514a43 autoconf/cmake: Always require isl code generation.
This change ensures that isl is only detected if it includes code generation
support. This allows us to remove a lot of conditional compilation and also
avoids missing test cases in case the feature is not available.

llvm-svn: 166403
2012-10-21 21:48:21 +00:00
Tobias Grosser c845fef391 Detect the isl code generation feature correctly
llvm-svn: 165034
2012-10-02 19:50:22 +00:00
Tobias Grosser 1df5289782 autoconf: Only define GPGPU_CODEGEN, if that feature is requested
Before we defined GPGPU_CODEGEN to '0', which does not disable the relevant code
as we just check if that value is defined at all. We now follow the cmake
approach and only define GPGPU_CODEGEN, if the feature should be enabled.

Reported by: Sebastian Pop <spop@codeaurora.org>

llvm-svn: 162275
2012-08-21 12:29:10 +00:00
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
Tobias Grosser c11349c55a Add support for libpluto as the scheduling optimizer.
llvm-svn: 161157
2012-08-02 07:47:26 +00:00
Tobias Grosser 88aeaf6ac4 Detect the cuda library available.
We will use the cuda library for the upcoming automatic GPGPU code generation.

Contributed by: Yabin Hu  <yabin.hwu@gmail.com>

llvm-svn: 158064
2012-06-06 12:16:10 +00:00
Sebastian Pop 082cea8616 add a check for ISL codegen at configure time
llvm-svn: 156305
2012-05-07 16:20:07 +00:00
Sebastian Pop c200977905 compile cloog code only when CLOOG_FOUND is set
llvm-svn: 156199
2012-05-04 20:30:03 +00:00
Tobias Grosser 70738df671 configure: Add gmp_inc when checking for CLooG
Otherwise configure fails if gmp is installed in a non default location.

llvm-svn: 141067
2011-10-04 06:55:03 +00:00
Tobias Grosser 758053788b Add initial version of Polly
This version is equivalent to commit ba26ebece8f5be84e9bd6315611d412af797147e
in the old git repository.

llvm-svn: 130476
2011-04-29 06:27:02 +00:00