niosega (Bonfante Nicolas)
User

Projects

User does not belong to any projects.

User Details

User Since
Apr 26 2017, 8:04 AM (33 w, 7 h)

Recent Activity

Aug 22 2017

niosega added a comment to D37010: [Polly][PM] Properly require and preservation of OptimizationRemarkEmitter..

This works well with MSE. Thanks a lot for this patch.

Aug 22 2017, 8:21 AM · Restricted Project

Aug 19 2017

niosega created D36926: [Polly][WIP] Clear DependenceInfo after expansion.
Aug 19 2017, 9:48 AM
niosega added inline comments to D36647: [Polly][WIP] Scalar fully indexed expansion.
Aug 19 2017, 8:10 AM
niosega updated the diff for D36647: [Polly][WIP] Scalar fully indexed expansion.

Add test cases for Phi expansion and one 'non-trivial' for array expansion.

Aug 19 2017, 8:05 AM

Aug 18 2017

niosega updated the diff for D36647: [Polly][WIP] Scalar fully indexed expansion.

Add a check that verify, before expansion (in isExpandable) if the expansion would lead to a read from the original value of the scalar.

Aug 18 2017, 10:30 AM
niosega added inline comments to D36647: [Polly][WIP] Scalar fully indexed expansion.
Aug 18 2017, 9:23 AM
niosega updated the diff for D36647: [Polly][WIP] Scalar fully indexed expansion.

Rebase with the new version of master to fix the bug of not correct dependencies (https://reviews.llvm.org/D36791).
Take Michael and Andreas comments into account.

Aug 18 2017, 9:17 AM
niosega added inline comments to D36791: [Polly][Bug fix] Wrong dependences filtering during Fully Indexed expansion.
Aug 18 2017, 8:09 AM
niosega added inline comments to D36791: [Polly][Bug fix] Wrong dependences filtering during Fully Indexed expansion.
Aug 18 2017, 7:46 AM
niosega added inline comments to D36791: [Polly][Bug fix] Wrong dependences filtering during Fully Indexed expansion.
Aug 18 2017, 6:52 AM
niosega updated the diff for D36791: [Polly][Bug fix] Wrong dependences filtering during Fully Indexed expansion.

Take Michael's suggestions into account.

Aug 18 2017, 6:51 AM

Aug 16 2017

niosega updated the diff for D36791: [Polly][Bug fix] Wrong dependences filtering during Fully Indexed expansion.

Take Andreas remarks into account.

Aug 16 2017, 10:27 AM
niosega created D36791: [Polly][Bug fix] Wrong dependences filtering during Fully Indexed expansion.
Aug 16 2017, 6:52 AM

Aug 12 2017

niosega created D36647: [Polly][WIP] Scalar fully indexed expansion.
Aug 12 2017, 1:50 PM

Aug 3 2017

niosega updated the summary of D34982: [Polly][WIP] Fully-Indexed static expansion.
Aug 3 2017, 3:23 PM
niosega updated the summary of D34982: [Polly][WIP] Fully-Indexed static expansion.
Aug 3 2017, 3:20 PM
niosega updated the diff for D34982: [Polly][WIP] Fully-Indexed static expansion.

Take into account Michaels comments.
Update setNewAccessRelation call (isl::map as parameter instead of isl map * )

Aug 3 2017, 3:18 PM

Jul 29 2017

niosega added inline comments to D34982: [Polly][WIP] Fully-Indexed static expansion.
Jul 29 2017, 2:03 PM

Jul 27 2017

niosega updated the diff for D34982: [Polly][WIP] Fully-Indexed static expansion.

Take Michael comments into account.

Jul 27 2017, 7:56 AM

Jul 26 2017

niosega added inline comments to D34982: [Polly][WIP] Fully-Indexed static expansion.
Jul 26 2017, 8:01 AM
niosega updated the diff for D34982: [Polly][WIP] Fully-Indexed static expansion.

Emit remarks instead of stderr printing. Test case works.

Jul 26 2017, 5:50 AM
niosega updated the diff for D34982: [Polly][WIP] Fully-Indexed static expansion.

Take into account Michael and Andreas comments.

Jul 26 2017, 4:39 AM

Jul 24 2017

niosega updated the diff for D34982: [Polly][WIP] Fully-Indexed static expansion.

Remove debug if condition.

Jul 24 2017, 12:52 PM
niosega updated the diff for D34982: [Polly][WIP] Fully-Indexed static expansion.

Take into account remarks from Michael.

Jul 24 2017, 12:46 PM

Jul 21 2017

niosega added inline comments to D34982: [Polly][WIP] Fully-Indexed static expansion.
Jul 21 2017, 9:18 AM

Jul 20 2017

niosega updated the diff for D34982: [Polly][WIP] Fully-Indexed static expansion.

In this revision, we have done :

Jul 20 2017, 7:47 AM

Jul 12 2017

niosega updated the diff for D34982: [Polly][WIP] Fully-Indexed static expansion.

Fully-Indexed expansion of the write accesses :

  • Build the new access map from the current access map
  • Create a new SAI for the expanded version of the access array or scalar
  • Modify the memory access to the new SAI
Jul 12 2017, 12:28 PM

Jul 5 2017

niosega added inline comments to D34982: [Polly][WIP] Fully-Indexed static expansion.
Jul 5 2017, 2:10 AM
niosega updated the summary of D34982: [Polly][WIP] Fully-Indexed static expansion.
Jul 5 2017, 2:04 AM

Jul 4 2017

niosega created D34982: [Polly][WIP] Fully-Indexed static expansion.
Jul 4 2017, 7:18 AM

Jun 28 2017

niosega updated the diff for D33688: [Polly] Heap allocation for new arrays.

There are no valid reasons not to use getElemSizeInBytes instead of getPrimitiveSizeInBits. This solution (with getElemSizeInBytes) is much more cleaner.

Jun 28 2017, 5:39 AM
niosega updated the summary of D33688: [Polly] Heap allocation for new arrays.
Jun 28 2017, 4:14 AM
niosega updated the diff for D33688: [Polly] Heap allocation for new arrays.

Rebase with master to take into account the commit of the check size for JSONImporter.

Jun 28 2017, 1:10 AM

Jun 27 2017

niosega updated the diff for D33688: [Polly] Heap allocation for new arrays.
Jun 27 2017, 1:29 PM
niosega updated the diff for D33688: [Polly] Heap allocation for new arrays.
  • Refactor code in JSONExporter for heap allocation detection.
  • Remove the test case concerning
  • Add CHECK in the test case for malloc/free to check if the inserting positions are correct.
Jun 27 2017, 1:11 PM
niosega updated the diff for D33688: [Polly] Heap allocation for new arrays.

This update modify two things :

  • Modifying the comment in CodeGeneration about the changes of insert point.
  • Remove the constness of CreateScopArrayInfo and getOrCreateScopArrayInfo and call directly the setter setIsOnHeap in JSONImporter.
Jun 27 2017, 7:13 AM
niosega added inline comments to D33688: [Polly] Heap allocation for new arrays.
Jun 27 2017, 4:51 AM

Jun 26 2017

niosega added inline comments to D33688: [Polly] Heap allocation for new arrays.
Jun 26 2017, 10:52 AM
niosega updated the diff for D33688: [Polly] Heap allocation for new arrays.

The changes made in this update are the following :

Jun 26 2017, 10:41 AM

Jun 23 2017

niosega added inline comments to D33688: [Polly] Heap allocation for new arrays.
Jun 23 2017, 2:27 AM

Jun 8 2017

niosega updated the diff for D33688: [Polly] Heap allocation for new arrays.

Release the memory.

Jun 8 2017, 9:04 AM

Jun 6 2017

niosega added a comment to D33688: [Polly] Heap allocation for new arrays.

I am wondering why you decided for a global stack-or-heap property. I'd assumed that this is a per-array decision. For instance, Roman's gemm optimization is designed to fit on the stack, there is no reason to allocate it on the heap.

Jun 6 2017, 2:43 PM
niosega updated the diff for D33688: [Polly] Heap allocation for new arrays.
Jun 6 2017, 9:13 AM
niosega updated the diff for D33688: [Polly] Heap allocation for new arrays.
Jun 6 2017, 8:41 AM
niosega retitled D33688: [Polly] Heap allocation for new arrays from Heap allocation for new arrays to [Polly] Heap allocation for new arrays.
Jun 6 2017, 8:40 AM

May 30 2017

niosega added a reviewer for D33688: [Polly] Heap allocation for new arrays: simbuerg.
May 30 2017, 12:42 PM
niosega created D33688: [Polly] Heap allocation for new arrays.
May 30 2017, 12:42 PM

May 23 2017

niosega updated the summary of D32739: [Polly] JSONImporter misses checks whether the data it imports makes sense..
May 23 2017, 8:55 AM
niosega updated the diff for D32739: [Polly] JSONImporter misses checks whether the data it imports makes sense..

Take into account the remarks of Tobias concerning the line before comments.
Also rebase with the last version of master.

May 23 2017, 7:29 AM

May 17 2017

niosega updated the diff for D32739: [Polly] JSONImporter misses checks whether the data it imports makes sense..
May 17 2017, 12:59 AM

May 15 2017

niosega updated the diff for D32739: [Polly] JSONImporter misses checks whether the data it imports makes sense..
May 15 2017, 4:20 AM

May 11 2017

niosega added a comment to D32739: [Polly] JSONImporter misses checks whether the data it imports makes sense..

I am currently working on adding checks for importAccess and cleaning the test cases (explain what they are doing, better naming .. )

May 11 2017, 12:47 AM
niosega added a comment to D32739: [Polly] JSONImporter misses checks whether the data it imports makes sense..
May 11 2017, 12:46 AM

May 3 2017

niosega updated the diff for D32739: [Polly] JSONImporter misses checks whether the data it imports makes sense..

This diff takes into account the remarks of Michael Kruse and Andreas Simbuerger.

May 3 2017, 7:45 AM

May 2 2017

niosega updated the diff for D32739: [Polly] JSONImporter misses checks whether the data it imports makes sense..

The first one is not complete.

May 2 2017, 4:43 AM
niosega updated the summary of D32739: [Polly] JSONImporter misses checks whether the data it imports makes sense..
May 2 2017, 3:54 AM
niosega created D32739: [Polly] JSONImporter misses checks whether the data it imports makes sense..
May 2 2017, 3:53 AM