Do Not Merge: Changes to DaCe CUDA compilation flags #2
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
commit 635da6c Author: Ioannis Magkanaris <iomagkanaris@gmail.com> Date: Fri Apr 25 15:40:25 2025 +0200 Add NVTX range in CUDA GPU kernel call of program
commit aef9945 Author: Philip Mueller <philip.mueller@cscs.ch> Date: Tue Mar 25 06:52:35 2025 +0100 As an experiment removed some code I think is useless, let's see what the tests say. commit e5bf87f Author: Philip Mueller <philip.mueller@cscs.ch> Date: Mon Mar 24 07:05:18 2025 +0100 Added a comment to address the possible issues with viewes. commit ba97874 Merge: b0b9945 4245396 Author: Philip Mueller <philip.mueller@cscs.ch> Date: Fri Mar 21 16:07:12 2025 +0100 Merge remote-tracking branch 'spcl/main' into improved-2d-copy commit b0b9945 Author: Philip Mueller <philip.mueller@cscs.ch> Date: Fri Mar 21 16:06:09 2025 +0100 Added Alexnicks's suggestions. commit 065e0d7 Author: Philip Mueller <philip.mueller@cscs.ch> Date: Mon Mar 17 08:56:39 2025 +0100 Added tests to ensure that the new verification works as expected. commit 51182e5 Author: Philip Mueller <philip.mueller@cscs.ch> Date: Mon Mar 17 08:14:14 2025 +0100 Moved the test for negative sized subsets from the Memlet to the `vaidate_state()` function. The reason is that in some cases this is valid, for example if an edge connects an AccessNode and a MapEntry, because, in that case the map might not be executed. Since the Memlet does not have access to its source and destination node it can not check that, so the test was moved to a location that can do this check. However, it only does the check for AN to AN connections, which is a bit restrictive, but this is something for later. commit 2801967 Author: Philip Mueller <philip.mueller@cscs.ch> Date: Mon Mar 17 07:41:51 2025 +0100 I am not sure why the printout of the edge is not correct, but it is not where I though I found it. commit 3166302 Author: Philip Mueller <philip.paul.mueller@bluemain.ch> Date: Sat Mar 15 08:23:21 2025 +0100 Fixed some issue and made it more logical. commit 02d87b5 Merge: 801adb1 d130792 Author: Philip Mueller <philip.paul.mueller@bluemain.ch> Date: Sat Mar 15 08:18:08 2025 +0100 Merge remote-tracking branch 'spcl/main' into improved-2d-copy commit 801adb1 Author: Philip Mueller <philip.paul.mueller@bluemain.ch> Date: Sat Mar 15 08:15:45 2025 +0100 Added more verification. commit 66b43f8 Author: Philip Mueller <philip.paul.mueller@bluemain.ch> Date: Sat Mar 15 07:57:03 2025 +0100 Simplified some check. commit 76a1a58 Author: Philip Mueller <philip.mueller@cscs.ch> Date: Fri Mar 14 15:20:49 2025 +0100 Added a new test for the pseudo 1d case, i.e. when we reduce a copy 2D copy to a 1d copy, because it happens to be continiously allocated. commit 0b15a74 Author: Philip Mueller <philip.mueller@cscs.ch> Date: Fri Mar 14 14:59:03 2025 +0100 Added a note about wrong usage of eid in validation. commit 322ecda Author: Philip Mueller <philip.mueller@cscs.ch> Date: Fri Mar 14 14:52:28 2025 +0100 Improved memlet checking. commit 61ea7a6 Author: Philip Mueller <philip.mueller@cscs.ch> Date: Fri Mar 14 14:43:27 2025 +0100 Added a new test to the SDFG. commit a67ad2a Author: Philip Mueller <philip.mueller@cscs.ch> Date: Fri Mar 14 14:36:02 2025 +0100 Added now also test for testing strided 1d copy. commit c931b91 Author: Philip Mueller <philip.mueller@cscs.ch> Date: Fri Mar 14 14:07:16 2025 +0100 Now 2D copies works, more tests needed. commit d0a396f Author: Philip Mueller <philip.mueller@cscs.ch> Date: Fri Mar 14 13:42:18 2025 +0100 Updated the memlet copying, I think I now have all the cases will now make the tests. commit 9b49c9e Author: Philip Mueller <philip.mueller@cscs.ch> Date: Fri Mar 14 11:16:12 2025 +0100 Made a first version of the new copy implementation. But I have to run the unit tests. commit feea97f Author: Philip Mueller <philip.mueller@cscs.ch> Date: Fri Mar 14 10:18:08 2025 +0100 Started with the implementation of a better copy, but I have to fix it more.
commit 5197dcf Author: Edoardo Paone <edoardo.paone@cscs.ch> Date: Wed Apr 30 07:43:10 2025 +0200 fix in cpp codegen commit c01760c Author: Edoardo Paone <edoardo.paone@cscs.ch> Date: Wed Apr 30 07:42:53 2025 +0200 Revert "add default - needed" This reverts commit 47d65ac. commit 580f743 Author: Edoardo Paone <edoardo.paone@cscs.ch> Date: Wed Apr 30 07:42:05 2025 +0200 Revert "edit" This reverts commit a7138d2. commit a7138d2 Author: Edoardo Paone <edoardo.paone@cscs.ch> Date: Wed Apr 30 07:26:14 2025 +0200 edit commit 47d65ac Author: Edoardo Paone <edoardo.paone@cscs.ch> Date: Tue Apr 29 23:27:55 2025 +0200 add default - needed commit a10b5b3 Author: Edoardo Paone <edoardo.paone@cscs.ch> Date: Tue Apr 29 23:07:09 2025 +0200 minor edit commit 40395a8 Author: Edoardo Paone <edoardo.paone@cscs.ch> Date: Tue Apr 29 22:42:22 2025 +0200 use ContextVar for _in_device_code
…CUDA compilation flags
philip-paul-mueller
changed the title
|
This is integrated into PR#1. |
philip-paul-mueller
force-pushed
the
gt4py-next-integration
branch
from
8b9b674 to
268fc18
Compare
philip-paul-mueller
force-pushed
the
gt4py-next-integration
branch
3 times, most recently
from
964e84b to
2d85437
Compare
philip-paul-mueller
force-pushed
the
gt4py-next-integration
branch
from
2d85437 to
9f72250
Compare
This was referenced Jun 5, 2025
|
Has been replaced with PR#4 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Added compilation flag for line info and removed fast_math flag from CUDA compilation flags