Repository navigation

Issues

New issue

cartesian: K-offset writes bound check is failing
#1754 · FlorianDeconinck opened on Nov 27, 2024
5

cartesian: computation_src cache in _make_extension_sources never hits (backend_data returns a copy)
GridTools/gt4py#2932 · opened on Oct 2, 2026
cyclic imports
GridTools/gt4py#2904 · opened on Sep 21, 2026
2
[next]: Length consistency of variable-length tuple arguments and 'out' is not checked
GridTools/gt4py#2903 · opened on Sep 21, 2026
[next]: Compiled backends cannot execute a program with an empty tuple output
GridTools/gt4py#2856 · opened on Sep 1, 2026
InlineDynamicShifts: let-bound producer shared between dynamically shifted consumers is not inlined
GridTools/gt4py#2839 · opened on Aug 27, 2026
next: ROCm process-pool compile workers probe the GPU for the device arch and can silently build for the wrong architecture (no HIPARCHS forwarding)
GridTools/gt4py#2802 · opened on Aug 19, 2026
1
next: a shifted read of a concat_where result is materialized in dace backend
GridTools/gt4py#2793 · opened on Aug 14, 2026
next[dace]: document indeterminism debug procedure
GridTools/gt4py#2781 · opened on Aug 13, 2026
next[dace]: subset_merger cannot decide coverage of overlapping reads
GridTools/gt4py#2776 · opened on Aug 11, 2026
1
next[dace]: track_view raises on a lowering produced View, and reconfigure_dataflow_after_rerouting asserts on an uninitialised Memlet
GridTools/gt4py#2775 · opened on Aug 11, 2026
5
next[dace]: inline_dataflow_into_map widens input Memlets to the full array
GridTools/gt4py#2774 · opened on Aug 10, 2026
3
cartesian: parser gets tripped by Field | None
GridTools/gt4py#2761 · opened on Aug 7, 2026
2
test_unlisted_construct_falls_back_to_ast_name does not exercise the fallback path
GridTools/gt4py#2759 · opened on Aug 6, 2026
Unsupported-syntax diagnostics are preempted by closure-variable resolution
GridTools/gt4py#2758 · opened on Aug 6, 2026
Domain inference extends scan domain for guarded offset access
GridTools/gt4py#2713 · opened on Jul 21, 2026
next: clarify who is responsible for gpu arch detection
GridTools/gt4py#2694 · opened on Jul 7, 2026
next[bug]: Validate dimension ordering on NdArrayField construction
GridTools/gt4py#2692 · opened on Jul 3, 2026
cartesian[dace]: SDFG validation warnings for read after write
GridTools/gt4py#2665 · opened on Jun 16, 2026
5
cartesian: report write-only temporaries
GridTools/gt4py#2618 · opened on Jun 1, 2026
1
next: Unify error messages in embedded and compiled backends
GridTools/gt4py#2617 · opened on Jun 1, 2026
next: naming of collections/named collections/tuples
GridTools/gt4py#2516 · opened on Mar 9, 2026
next: builtin function type hints block client type checking
GridTools/gt4py#2507 · opened on Mar 2, 2026
cartesian: [CI Failure | UTests] Caching test using build time check is not stable
Bug
GridTools/gt4py#2466 · opened on Jan 30, 2026
cartesian: Daily CI failures with MacOS + py > 3.11 + uv "highest" strategy.
Bug
GridTools/gt4py#2439 · opened on Jan 15, 2026
next: make it possible ot have caches on readonly fs
GridTools/gt4py#2336 · opened on Oct 23, 2025

Navigation Menu

Repository navigation

Issues

Issues

cartesian: `computation_src` cache in `_make_extension_sources` never hits (`backend_data` returns a copy)

cyclic imports

[next]: Length consistency of variable-length tuple arguments and 'out' is not checked

[next]: Compiled backends cannot execute a program with an empty tuple output

InlineDynamicShifts: let-bound producer shared between dynamically shifted consumers is not inlined

next: ROCm process-pool compile workers probe the GPU for the device arch and can silently build for the wrong architecture (no HIPARCHS forwarding)

next: a shifted read of a concat_where result is materialized in dace backend

next[dace]: document indeterminism debug procedure

next[dace]: `subset_merger` cannot decide coverage of overlapping reads

next[dace]: `track_view` raises on a lowering produced View, and `reconfigure_dataflow_after_rerouting` asserts on an uninitialised Memlet

next[dace]: `inline_dataflow_into_map` widens input Memlets to the full array

cartesian: parser gets tripped by `Field | None`

test_unlisted_construct_falls_back_to_ast_name does not exercise the fallback path

Unsupported-syntax diagnostics are preempted by closure-variable resolution

Domain inference extends scan domain for guarded offset access

next: clarify who is responsible for gpu arch detection

next[bug]: Validate dimension ordering on `NdArrayField` construction

cartesian[dace]: SDFG validation warnings for read after write

cartesian: report write-only temporaries

next: Unify error messages in embedded and compiled backends

next: naming of collections/named collections/tuples

next: builtin function type hints block client type checking

cartesian: [CI Failure | UTests] Caching test using build time check is not stable

cartesian: Daily CI failures with MacOS + py > 3.11 + uv "highest" strategy.

next: make it possible ot have caches on readonly fs