{"payload":{"feedbackUrl":"https://github.com/orgs/community/discussions/53140","repo":{"id":219703233,"defaultBranch":"main","name":"gt4py","ownerLogin":"GridTools","currentUserCanPush":false,"isFork":false,"isEmpty":false,"createdAt":"2019-11-05T09:12:36.000Z","ownerAvatar":"https://avatars.githubusercontent.com/u/18595324?v=4","public":true,"private":false,"isOrgOwned":true},"refInfo":{"name":"","listCacheKey":"v0:1726838736.0","currentOid":""},"activityList":{"items":[{"before":"21b1dfc8a80ffd9f1825c1c72ca34c25e3db3c4c","after":"b6d44277ab6200c2371ac19e72397ada2e969886","ref":"refs/heads/main","pushedAt":"2024-09-20T13:23:20.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"havogt","name":"Hannes Vogt","path":"/havogt","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/5136069?s=80&v=4"},"commit":{"message":"Releasing 1.0.4 version. (#1654)","shortMessageHtmlLink":"Releasing 1.0.4 version. (#1654)"}},{"before":"1aba51a86f3ba9196cb2d9e53f9b1b760e936ef2","after":"eaa06ec9af8dd27bcd702609f1e31d1c15005296","ref":"refs/heads/enable_gh_ci","pushedAt":"2024-09-20T12:35:01.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"havogt","name":"Hannes Vogt","path":"/havogt","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/5136069?s=80&v=4"},"commit":{"message":"Update cscs-ci.yml","shortMessageHtmlLink":"Update cscs-ci.yml"}},{"before":null,"after":"1aba51a86f3ba9196cb2d9e53f9b1b760e936ef2","ref":"refs/heads/enable_gh_ci","pushedAt":"2024-09-20T11:51:21.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"havogt","name":"Hannes Vogt","path":"/havogt","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/5136069?s=80&v=4"},"commit":{"message":"Re-enable CI on GH200","shortMessageHtmlLink":"Re-enable CI on GH200"}},{"before":"fe6dbd4a300b1c88fd90aca806ed431a70977e04","after":"21b1dfc8a80ffd9f1825c1c72ca34c25e3db3c4c","ref":"refs/heads/main","pushedAt":"2024-09-20T07:07:53.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"tehrengruber","name":"Till Ehrengruber","path":"/tehrengruber","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6791445?s=80&v=4"},"commit":{"message":"feat[next]: domain inference for let, make_tuple, tuple_get, cond (#1591)\n\nInfers the minimal domain of (nested) `let`, `make_tuple`, `tuple_get`,\r\n`cond` and other builtins as an extension to PR #1568\r\n\r\n- New functions `infer_let`, `infer_make_tuple`, `infer_tuple_get`,\r\n`infer_cond` in `gt4py.next.iterator.transforms.infer_domain`\r\n- New function `infer_expr` in\r\ngt4py.next.iterator.transforms.infer_domain which calls the appropriate\r\nof the above (or `infer_as_fieldop` and `infer_program`)\r\n- Several new tests in test_infer_domain.py to test functionality\r\n\r\nNote:\r\nTemporary handling was only present until commit fc4846f and has been\r\nremoved in commit e8e679d to reduce unneeded complexity. This pass will\r\nbe executed before temporary extraction, hence there exist valid\r\n`domain`s in all program calls, i.e. all `SetAt` do have a domain (not\r\n`AUTO_DOMAIN`) that doesn't need to be inferred.\r\n\r\n---------\r\n\r\nCo-authored-by: Till Ehrengruber ","shortMessageHtmlLink":"feat[next]: domain inference for let, make_tuple, tuple_get, cond (#1591"}},{"before":"2eed3679b513c5e308fea7a666f9a3b6bb497ec6","after":"fe6dbd4a300b1c88fd90aca806ed431a70977e04","ref":"refs/heads/main","pushedAt":"2024-09-20T06:56:33.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"tehrengruber","name":"Till Ehrengruber","path":"/tehrengruber","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6791445?s=80&v=4"},"commit":{"message":"bug[next]: Enable test_ffront_lap again (#1652)","shortMessageHtmlLink":"bug[next]: Enable test_ffront_lap again (#1652)"}},{"before":"789a5848e8fbf6ac43ab2337820f99005e7c9bf3","after":"2eed3679b513c5e308fea7a666f9a3b6bb497ec6","ref":"refs/heads/main","pushedAt":"2024-09-19T13:15:08.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"philip-paul-mueller","name":"Philip Müller","path":"/philip-paul-mueller","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/147368808?s=80&v=4"},"commit":{"message":"feat[dace]: Custom SDFG inline pass (#1649)\n\nAdded a custom pass for inlining SDFG. The function builds upon the traditional inline pass (`InlineSDFG`), however, before it is run some cleaning steps are preformed (`PruneConectors` and `PruneSymbols`) which increase the likelihood that the inlining can be done. The cost is that state fusing is performed.\r\n\r\nAlso the `gt_simplify()` function is modified, instead of the build in behaviour it will only run the GT4Py specific one.\r\nThis behaviour can not be changed.","shortMessageHtmlLink":"feat[dace]: Custom SDFG inline pass (#1649)"}},{"before":null,"after":"7c5c85d467f04e4a2436279be1f6178d2ef93831","ref":"refs/heads/enable_lap_test_again","pushedAt":"2024-09-19T12:49:45.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"tehrengruber","name":"Till Ehrengruber","path":"/tehrengruber","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6791445?s=80&v=4"},"commit":{"message":"Enable test_ffront_lap again","shortMessageHtmlLink":"Enable test_ffront_lap again"}},{"before":"acefd84b56c9d1af6c317ae853e58554a4bb57c5","after":"789a5848e8fbf6ac43ab2337820f99005e7c9bf3","ref":"refs/heads/main","pushedAt":"2024-09-19T10:27:43.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"edopao","name":null,"path":"/edopao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/25990065?s=80&v=4"},"commit":{"message":"feat[next][dace]: add workflow for dace fieldview backend (#1623)\n\nThis PR includes some code changes from `gtir` integration branch\r\nrelative to the setup of the dace backend. It tries to share code which\r\nis common between the `dace_iterator` and the `dace_fieldview` backends.","shortMessageHtmlLink":"feat[next][dace]: add workflow for dace fieldview backend (#1623)"}},{"before":"07e2ee9e2906a3eab32dc651fecc7ee63340aefc","after":"acefd84b56c9d1af6c317ae853e58554a4bb57c5","ref":"refs/heads/main","pushedAt":"2024-09-19T10:26:41.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"edopao","name":null,"path":"/edopao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/25990065?s=80&v=4"},"commit":{"message":"feat[next][dace]: generic scalar expressions as fieldop args (#1633)\n\nThis PR adds support for symbolic expressions as field-operator\r\narguments. This is needed to support expressions like `(0 - 4)` which is\r\nused in GTIR to represent the negative value `-4`. It also supports\r\ngeneric scalar expressions, e.g. `cast_(tuple_get(0, var), \"float64\")`.","shortMessageHtmlLink":"feat[next][dace]: generic scalar expressions as fieldop args (#1633)"}},{"before":"08e063ca1512cc5af346b0f31d3bc23cf05c63a3","after":"07e2ee9e2906a3eab32dc651fecc7ee63340aefc","ref":"refs/heads/main","pushedAt":"2024-09-18T10:40:37.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"philip-paul-mueller","name":"Philip Müller","path":"/philip-paul-mueller","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/147368808?s=80&v=4"},"commit":{"message":"feat[dace]: Modified `gt_simplify()` (#1647)\n\nBefore the pass was just calling the native DaCe version. Now it will\r\nnot call the `PromoteScalarToSymbol` and `ConstantPropagation` pass.\r\nThis is because the lowering sometimes has to change between a symbol\r\nand a scalar and back and back again and so on. Furthermore, it looks\r\nlike these passes have problems with that, so we excluded them.\r\n\r\nThis is a temporary solution, at the end, it might be feasible or good\r\nto run the full simplify pass.","shortMessageHtmlLink":"feat[dace]: Modified gt_simplify() (#1647)"}},{"before":"9328c5069bad4ab292c9c9853dba0d83a59dc17b","after":"08e063ca1512cc5af346b0f31d3bc23cf05c63a3","ref":"refs/heads/main","pushedAt":"2024-09-18T08:22:25.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"egparedes","name":"Enrique González Paredes","path":"/egparedes","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/18477?s=80&v=4"},"commit":{"message":"chore: add pre-commit hook for the tach tool to control dependencies (#1641)\n\nThis PR only adds the configuration for the\r\n[`tach`](https://docs.gauge.sh/getting-started/introduction) tool and\r\nits pre-commit hook in non-strict mode. In further PRs stricter checks\r\ncan be added.\r\n\r\nCo-authored-by: Hannes Vogt ","shortMessageHtmlLink":"chore: add pre-commit hook for the tach tool to control dependencies (#…"}},{"before":"c8822c05d5097734b2e457e11b314c5487a9d94b","after":"9328c5069bad4ab292c9c9853dba0d83a59dc17b","ref":"refs/heads/main","pushedAt":"2024-09-17T17:47:47.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"tehrengruber","name":"Till Ehrengruber","path":"/tehrengruber","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6791445?s=80&v=4"},"commit":{"message":"feat[next]: Refactor CSE pass to support ITIR.Program (#1646)\n\nExtends the common subexpression elimination to support the new\r\nitir.Program node and pushes the intermediate Fencil -> Program\r\nconversion upwards the pass manager. The CSE pass now uses the type\r\ninference such that only field expressions or composites thereof are\r\ncollected in field-view context (i.e. outside of as_fieldop).\r\n\r\nThis PR was initially meant to be merged into the temporary GTIR branch\r\nand reviewed by @egparedes here: #1570. The only change since then is to\r\nmake dace tests pass (see commit 160a616).\r\n\r\n---------\r\n\r\nCo-authored-by: edopao \r\nCo-authored-by: Enrique González Paredes ","shortMessageHtmlLink":"feat[next]: Refactor CSE pass to support ITIR.Program (#1646)"}},{"before":"394868a28f16ca74d4e59648a0efb67c21b3adbc","after":"c8822c05d5097734b2e457e11b314c5487a9d94b","ref":"refs/heads/main","pushedAt":"2024-09-17T16:43:13.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"tehrengruber","name":"Till Ehrengruber","path":"/tehrengruber","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6791445?s=80&v=4"},"commit":{"message":"feat[next]: Prepare TraceShift pass for GTIR (#1592)\n\nThe pass now only operates on stencils instead of `StencilClosure`s\r\nusing the newly introduced function `trace_stencil`.","shortMessageHtmlLink":"feat[next]: Prepare TraceShift pass for GTIR (#1592)"}},{"before":"15baa374178b89ae81b7d333a10f2cdf30ea19d1","after":"394868a28f16ca74d4e59648a0efb67c21b3adbc","ref":"refs/heads/main","pushedAt":"2024-09-17T11:29:30.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"havogt","name":"Hannes Vogt","path":"/havogt","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/5136069?s=80&v=4"},"commit":{"message":"bug[next]: fix to gtir lowering of `astype` on tuples containing scalars (#1642)\n\nThe unpacked tuple will contain `as_fieldop`ed `cast_`s for fields, but\r\ndirect `cast_`s for scalars.\r\n\r\nNote: A similar change is not needed for `where` which also works on\r\ntuples in the `true` and `false` branches. The reason is that `mask` is\r\nrequired to be a `Field`, therefore `true` and `false` are promoted to\r\n`Field`s, too.","shortMessageHtmlLink":"bug[next]: fix to gtir lowering of astype on tuples containing scal…"}},{"before":"723d7d56acf67b1a419a27d81afa1870ba7d0710","after":"15baa374178b89ae81b7d333a10f2cdf30ea19d1","ref":"refs/heads/main","pushedAt":"2024-09-16T22:43:19.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"tehrengruber","name":"Till Ehrengruber","path":"/tehrengruber","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6791445?s=80&v=4"},"commit":{"message":"feature[next]: Nested scalars args & cleanup (#1540)\n\n- Cleanup handling of tuple arguments in GTFN backend. This enables\r\nusing scalar arguments in both the domain expression of a program and as\r\nan argument to a field operator simultaneously. Scalar arguments to an\r\napplied `as_fieldop` are detected using the new type inference.\r\n- Adds support for nested scalar args in the GTFN backend\r\n- Adds support for nested scalar & field args in GTFN and Embedded","shortMessageHtmlLink":"feature[next]: Nested scalars args & cleanup (#1540)"}},{"before":"0ddddd37d3056ad6518f33908eb02f3b1f992878","after":"723d7d56acf67b1a419a27d81afa1870ba7d0710","ref":"refs/heads/main","pushedAt":"2024-09-16T11:50:25.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"tehrengruber","name":"Till Ehrengruber","path":"/tehrengruber","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6791445?s=80&v=4"},"commit":{"message":"feat[next]: Type inference: Add support for tuple arguments in `cond` (#1626)","shortMessageHtmlLink":"feat[next]: Type inference: Add support for tuple arguments in cond (…"}},{"before":"cd20826e0da2e69c96ea6808453009ef7924c954","after":"0ddddd37d3056ad6518f33908eb02f3b1f992878","ref":"refs/heads/main","pushedAt":"2024-09-11T14:38:38.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"nfarabullini","name":"Nicoletta Farabullini","path":"/nfarabullini","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/41536517?s=80&v=4"},"commit":{"message":"changed organization from C2SM to EXCLAIM in authors (#1640)\n\nfrom:\r\n```\r\nFarabullini, Nicoletta. ETH Zurich - C2SM\r\n```\r\n\r\nto:\r\n```\r\nFarabullini, Nicoletta. ETH Zurich - EXCLAIM\r\n```\r\nin Authors.md","shortMessageHtmlLink":"changed organization from C2SM to EXCLAIM in authors (#1640)"}},{"before":"712324cbf1c55b1113357b497975152a4af89646","after":"cd20826e0da2e69c96ea6808453009ef7924c954","ref":"refs/heads/main","pushedAt":"2024-09-11T11:28:33.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"tehrengruber","name":"Till Ehrengruber","path":"/tehrengruber","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6791445?s=80&v=4"},"commit":{"message":"bug[next]: Fix implicit offset provider (#1638)\n\n#1484 introduced support for writing `field(I+1)`. The implicit offset providers we generate for this accidentally also included local dimensions, which broke e.g. Icon4Py with errors like:\r\n```\r\n if not dim.kind == common.DimensionKind.VERTICAL:\r\n> raise ValueError(\r\n \"Mapping an offset to a horizontal dimension in unstructured is not allowed.\"\r\n )\r\nE ValueError: Mapping an offset to a horizontal dimension in unstructured is not allowed.\r\n```\r\nThis PR fixes that by only including horizontal and vertical, but not local dimensions in the implicit offset providers.","shortMessageHtmlLink":"bug[next]: Fix implicit offset provider (#1638)"}},{"before":"e0fb2a2410909a9ce63d0c15f33325dc3577eb70","after":"712324cbf1c55b1113357b497975152a4af89646","ref":"refs/heads/main","pushedAt":"2024-09-11T08:58:20.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"egparedes","name":"Enrique González Paredes","path":"/egparedes","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/18477?s=80&v=4"},"commit":{"message":"style: Remove stale references to isort (#1631)\n\n`isort` has been replaced by `ruff` in PR\r\nhttps://github.com/GridTools/gt4py/pull/1466.\r\n\r\n_Note_ Removing `isort` means rebuilding the requirements, which updates\r\na couple of unrelated (to the `isort` removal) dependencies.\r\n\r\n_Note 2_ I also checked for `black` (formatter) and there are lingering\r\nreferences too. However, `black` is actively used as formatter in code\r\ngeneration\r\n\r\n\r\nhttps://github.com/GridTools/gt4py/blob/70dd7c34036e48c37f5047bf32573916836a4a35/src/gt4py/eve/codegen.py#L106\r\n\r\nCo-authored-by: Roman Cattaneo <>","shortMessageHtmlLink":"style: Remove stale references to isort (#1631)"}},{"before":"fe8349e9e51f5aea9dfed3ebb88c6c11df4cbc06","after":"e0fb2a2410909a9ce63d0c15f33325dc3577eb70","ref":"refs/heads/main","pushedAt":"2024-09-11T08:52:57.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"egparedes","name":"Enrique González Paredes","path":"/egparedes","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/18477?s=80&v=4"},"commit":{"message":"refactor[cartesian]: readability improvements in gtir -> oir conversion and other cleanups (#1630)\n\nBeing new to the codebase and in preparation of the [GT4Py-DaCe bridge\r\nrefactor](https://github.com/GEOS-ESM/NDSL/issues/53) I was reading a\r\nlot of code. Some things that caught my eye are summarized in this PR.\r\nBigger changes include\r\n\r\n1. Moving `daceir.py` into the `dace/` backend folder. This follows the\r\nconvention of all other backends who have their IR inside the backend\r\nfolder. (last commit)\r\n2. Readability improvements in `gtir_to_oir.py` (the \"Cleanup visit_*\"\r\ncommits): In my opinion, there's no point in shortening \"statement\" to\r\n\"stmt\". I wouldn't go as far as renaming the classes. For the renamed\r\nlocal variables, I think this adds a lot to readability.\r\n3. `visit_While` in `gtir_to_oir.py` was having an extra mask statement\r\naround the `while` loop (in case a mask was defined. I inlined the\r\npotential `mask` with the condition of `while` loop.\r\n4. In `gtir_to_oir.py` don't translate every body statement into a\r\nsingle `MaskStmt`. Instead, create one (or two incase of `else`)\r\n`MaskStmt` with multiple statements in the `body` (which is already\r\ntyped as a list of statements).\r\n\r\nTested locally by running the test suite. All tests passing.\r\n\r\nCo-authored-by: Roman Cattaneo <>","shortMessageHtmlLink":"refactor[cartesian]: readability improvements in gtir -> oir conversi…"}},{"before":"03f100c69b6288f560466e576f486c5a50c1d1e5","after":"fe8349e9e51f5aea9dfed3ebb88c6c11df4cbc06","ref":"refs/heads/main","pushedAt":"2024-09-11T08:04:04.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"egparedes","name":"Enrique González Paredes","path":"/egparedes","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/18477?s=80&v=4"},"commit":{"message":"refactor: use modern-style type hints (#1632)\n\nIn the past, the only way to create type hints for yet-to-be-defined\r\ntypes was to quote them. Since python 3.7 we can get rid of the quotes\r\nwhen using `annotations` from `__future__` instead. This is supposed to\r\nbecome the default in the future.\r\n\r\nThis PR is a spin-off from\r\nhttps://github.com/GridTools/gt4py/pull/1630#issuecomment-2333757200.\r\n\r\nCo-authored-by: Roman Cattaneo <>","shortMessageHtmlLink":"refactor: use modern-style type hints (#1632)"}},{"before":"7f2c0d2dc435264d24f0e50072d33bb8071d351d","after":"03f100c69b6288f560466e576f486c5a50c1d1e5","ref":"refs/heads/main","pushedAt":"2024-09-11T06:39:48.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"edopao","name":null,"path":"/edopao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/25990065?s=80&v=4"},"commit":{"message":"test[dace]: Fix skip rules for dace optional dependency (#1637)\n\nThis PR fixes the solution to skip testcases dependent on dace provided\r\nin #1635. The problem was that `pytest.skip` function was called within\r\n`__init__` and/or `conftest` files in the test folder.\r\n\r\nAdditional changes:\r\n * Moved all dace-related tests in `dace_tests` package folder.\r\n * Introduced pytest marker `requires_dace` at package level, which needs\r\nto be included in all test files.\r\n * Run ruff on test code in `dace_tests` package to cleanup and sort the\r\nimport list.","shortMessageHtmlLink":"test[dace]: Fix skip rules for dace optional dependency (#1637)"}},{"before":"85893877e728bd4e6a58107f89aeb434ba29e921","after":"7f2c0d2dc435264d24f0e50072d33bb8071d351d","ref":"refs/heads/main","pushedAt":"2024-09-10T20:13:56.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"tehrengruber","name":"Till Ehrengruber","path":"/tehrengruber","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6791445?s=80&v=4"},"commit":{"message":"feature[next]: Extend cartesian offset syntax (#1484)\n\nAdd support for short-hand Cartesian syntax, e.g., `field(I+1,J-1)`.\r\n---------\r\n\r\nCo-authored-by: Till Ehrengruber ","shortMessageHtmlLink":"feature[next]: Extend cartesian offset syntax (#1484)"}},{"before":"42636f717aeebf248df163db17151cd6d5137e36","after":"85893877e728bd4e6a58107f89aeb434ba29e921","ref":"refs/heads/main","pushedAt":"2024-09-09T13:00:32.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"egparedes","name":"Enrique González Paredes","path":"/egparedes","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/18477?s=80&v=4"},"commit":{"message":" bug[next]: enable scalar math operations in embedded (#1636)\n\nRe-enable scalar math operations in embedded. This PR implements the\r\nbare minimum to enable the functionality, but more work should come in\r\nfuture PRs to enable proper unit testing and renaming some outdated\r\nbuiltins.\r\n\r\n---------\r\n\r\nCo-authored-by: Hannes Vogt ","shortMessageHtmlLink":" bug[next]: enable scalar math operations in embedded (#1636)"}},{"before":"70dd7c34036e48c37f5047bf32573916836a4a35","after":"42636f717aeebf248df163db17151cd6d5137e36","ref":"refs/heads/main","pushedAt":"2024-09-09T11:49:16.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"edopao","name":null,"path":"/edopao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/25990065?s=80&v=4"},"commit":{"message":"test[dace]: Add skip rules for dace optional dependency (#1635)\n\nTests that depend on dace should check if dace is installed, because\r\ndace is an optional dependency. Error detected in spack-c2sm CI where\r\ndace is not installed.\r\nAdditionally, added `requires_dace` marker to filter the testcases.","shortMessageHtmlLink":"test[dace]: Add skip rules for dace optional dependency (#1635)"}},{"before":"20082e93d644b5c258db0b5e82701b4956e5f36f","after":"70dd7c34036e48c37f5047bf32573916836a4a35","ref":"refs/heads/main","pushedAt":"2024-09-06T04:58:17.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"philip-paul-mueller","name":"Philip Müller","path":"/philip-paul-mueller","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/147368808?s=80&v=4"},"commit":{"message":"feat[next][dace]: Dace fieldview transformations (#1594)\n\nInitial PR for the Optimization pipeline of the DaCe backend in GT4Py.\r\nThe PR is very basic, but contains the basis for later advancement.\r\nEssentially, it defines the `gt_auto_optimize()` function which performs\r\noptimization on a basic heuristic.\r\n\r\nThere are some parts missing, such as:\r\n- There is no way to guide the optimization, i.e. a decision making\r\nprocess that determine if a map should be fused or not.\r\n- There is also no transformation which fuses parallel maps (was not\r\nneeded for this kernel).\r\n\r\n---------\r\n\r\nCo-authored-by: Edoardo Paone \r\nCo-authored-by: Philip Mueller \r\nCo-authored-by: Enrique González Paredes ","shortMessageHtmlLink":"feat[next][dace]: Dace fieldview transformations (#1594)"}},{"before":"07f37fbb41ddfae81f69ead674565d76120d0911","after":"20082e93d644b5c258db0b5e82701b4956e5f36f","ref":"refs/heads/main","pushedAt":"2024-09-05T13:15:54.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"FlorianDeconinck","name":"Florian Deconinck","path":"/FlorianDeconinck","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/5053496?s=80&v=4"},"commit":{"message":"fix[cartesian]: Fix serialize default behavior when Pickled property was not saved (#1629)\n\n## Description\r\n\r\nDaCe has a behavior of _not_ saving properties if they have a default\r\n(newish behavior) which leads our 2-step process of library node caching\r\nto fail. Since we have decided to rely on pickling to cache out the\r\ncontent of the LibraryNode we respond to this change in DaCe by making\r\nsure default values pass to the pickle deserializer are returned plain\r\n\r\n## Requirements\r\n\r\n- [ ] All fixes and/or new features come with corresponding tests.\r\n- [ ] Important design decisions have been documented in the approriate\r\nADR inside the [docs/development/ADRs/](docs/development/ADRs/Index.md)\r\nfolder.\r\n\r\n---------\r\n\r\nCo-authored-by: Florian Deconinck \r\nCo-authored-by: Roman Cattaneo ","shortMessageHtmlLink":"fix[cartesian]: Fix serialize default behavior when Pickled property …"}},{"before":"829b2eb8744261023bfc8c15f728044cbe03d9db","after":"07f37fbb41ddfae81f69ead674565d76120d0911","ref":"refs/heads/main","pushedAt":"2024-09-04T16:09:39.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"egparedes","name":"Enrique González Paredes","path":"/egparedes","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/18477?s=80&v=4"},"commit":{"message":"refactor[cartesian]: Enable MaskInlining in gt4yp - DaCe bridge (#1624)\n\n`MaskInlining` will be a key feature of the improved gt4py / DaCe bridge\r\nas envisioned in https://github.com/GEOS-ESM/NDSL/issues/53, which will\r\npropagate more control flow elements from gt4py to DaCe. We thus want to\r\nmake sure if `MaskInlining` itself is posing any problems. We tracked\r\nthe git history of this explicit exclude back to PR\r\nhttps://github.com/GridTools/gt4py/pull/300, which not only added the\r\nnpir backend, but also refactored how pipeline stages are defined.\r\nPrevious to this PR optimization passes to be specified explicitly and\r\nthe PR changed this to an exclude list. It looks like the `MaskInlining`\r\npass was added after the list of optimization passes was formed and thus\r\nexcluded (to keep the list constant). From this history we thus don't\r\nexpect `MaskInlining` to break anything.\r\n\r\nCo-authored-by: Roman Cattaneo <>","shortMessageHtmlLink":"refactor[cartesian]: Enable MaskInlining in gt4yp - DaCe bridge (#1624)"}},{"before":"feda23dfca150339158488591d4394efc9c3c968","after":"829b2eb8744261023bfc8c15f728044cbe03d9db","ref":"refs/heads/main","pushedAt":"2024-09-04T16:07:19.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"egparedes","name":"Enrique González Paredes","path":"/egparedes","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/18477?s=80&v=4"},"commit":{"message":"docs: Fixed typos in ADRs and a docstring (#1627)\n\nI noticed a couple of typos in the ADRs (and in one docstring) while I\r\nwas familiarizing myself with the codebase. Thanks for keeping an ADR\r\nand good documentation practices. Really helps if you are new to a\r\ncodebase.\r\n\r\nThis PR just fixes mentioned typos.\r\n\r\nCo-authored-by: Roman Cattaneo <>","shortMessageHtmlLink":"docs: Fixed typos in ADRs and a docstring (#1627)"}},{"before":"07e1b1d52aa40649eb48b20629a6b665424e4646","after":"feda23dfca150339158488591d4394efc9c3c968","ref":"refs/heads/main","pushedAt":"2024-09-04T12:23:26.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"edopao","name":null,"path":"/edopao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/25990065?s=80&v=4"},"commit":{"message":"feat[next][dace]: Build clean nestedSDFG without unused data connectors (#1628)\n\nThe main objective of this PR is to reduce the GT4Py dependency on DaCe\r\n(especially on DaCe transformations) to produce clean SDFGs. Instead of\r\nrelying on the `PruneConnectors` pass to remove unused data connectors\r\nin nested SDFGs, these connectors can be avoided in the first place,\r\nwhen lowering let-lambdas to nested SDFGs.","shortMessageHtmlLink":"feat[next][dace]: Build clean nestedSDFG without unused data connecto…"}}],"hasNextPage":true,"hasPreviousPage":false,"activityType":"all","actor":null,"timePeriod":"all","sort":"DESC","perPage":30,"cursor":"Y3Vyc29yOnYyOpK7MjAyNC0wOS0yMFQxMzoyMzoyMC4wMDAwMDBazwAAAAS7_hc4","startCursor":"Y3Vyc29yOnYyOpK7MjAyNC0wOS0yMFQxMzoyMzoyMC4wMDAwMDBazwAAAAS7_hc4","endCursor":"Y3Vyc29yOnYyOpK7MjAyNC0wOS0wNFQxMjoyMzoyNi4wMDAwMDBazwAAAASsrLd-"}},"title":"Activity · GridTools/gt4py"}