List view

Finish decoupling from upstream Numba

P0Blocking external stakeholdersBlocking external stakeholders

Status: Open.

Task

#778 In NVIDIA/numba-cuda;

· rparolin opened on Feb 5, 2026

[EPIC] Support lower precision data types

feature requestNew feature or requestNew feature or request

Status: Open.

Enhancement

#197 In NVIDIA/numba-cuda;

· kkraus14 opened on Apr 11, 2025

[FEA] Support fp8 data types

feature requestNew feature or requestNew feature or request

Status: Open.

#200 In NVIDIA/numba-cuda;

· kkraus14 opened on Apr 11, 2025

[FEA] Support fp6 data types

feature requestNew feature or requestNew feature or request

Status: Open.

#201 In NVIDIA/numba-cuda;

· kkraus14 opened on Apr 11, 2025

[FEA] Support fp4 data types

feature requestNew feature or requestNew feature or request

Status: Open.

#202 In NVIDIA/numba-cuda;

· kkraus14 opened on Apr 11, 2025

[FEA] Supporting DLPacks

feature requestNew feature or requestNew feature or request

Status: Open.

#122 In NVIDIA/numba-cuda;

· isVoid opened on Feb 7, 2025

[BUG] C ABI passes struct values incorrectly when compiling with <code>compile()</code>

bugSomething isn't workingSomething isn't working

Status: Open.

#175 In NVIDIA/numba-cuda;

· gmarkall opened on Mar 25, 2025

[BUG] Large overhead when launching kernel with torch tensors

bugSomething isn't workingSomething isn't working

Status: Open.

Bug

#98 In NVIDIA/numba-cuda;

· gmarkall opened on Dec 30, 2024

[FEA] Support for <code>__dlpack_capi__</code>

feature requestNew feature or requestNew feature or request

Status: Open.

#531 In NVIDIA/numba-cuda;

· gmarkall opened on Oct 16, 2025

[FEA] Lower precision data type inference

feature requestNew feature or requestNew feature or request

Status: Open.

Enhancement

#496 In NVIDIA/numba-cuda;

· rparolin opened on Oct 1, 2025

split up objects in <code>cudadrv/driver.py</code>

improvementImproves an existing functionalityImproves an existing functionality

Status: Open.

Enhancement

#645 In NVIDIA/numba-cuda;

· cpcloud opened on Dec 10, 2025

[BUG] _assign_kernel cache not invalidated after ctx.reset()

bugSomething isn't workingSomething isn't working

Status: Open.

#644 In NVIDIA/numba-cuda;

· shaunc opened on Dec 9, 2025

[RFC] Make <code>cuda.core</code> a required dependency

feature requestNew feature or requestNew feature or request

Status: Open.

Enhancement

#128 In NVIDIA/numba-cuda;

· leofang opened on Feb 20, 2025

Set up CI testing against Numba <code>main</code>

Status: Open.

#213 In NVIDIA/numba-cuda;

· kkraus14 opened on Apr 14, 2025

Remove CUDA-related items from upstream Numba

Status: Open.

#220 In NVIDIA/numba-cuda;

· kkraus14 opened on Apr 14, 2025

[FEA] Coverage reporting

feature requestNew feature or requestNew feature or request

Status: Open.

#346 In NVIDIA/numba-cuda;

· gmarkall opened on Jul 23, 2025

[TASK] Post-CUDA 13 driver fixups

Status: Open.

#386 In NVIDIA/numba-cuda;

· gmarkall opened on Aug 7, 2025

[BUG] Raising an exception or asserting inside a kernel is silently ignored.

bugSomething isn't workingSomething isn't working

Status: Open.

#388 In NVIDIA/numba-cuda;

· brycelelbach opened on Aug 7, 2025

Switch to use CuPy arrays in the test suite

Status: Open.

#472 In NVIDIA/numba-cuda;

· leofang opened on Sep 12, 2025

Reduce kernel launch overhead

Status: Open.

Epic

#477 In NVIDIA/numba-cuda;

· leofang opened on Sep 17, 2025

Support <code>ml_dtypes</code>

feature requestNew feature or requestNew feature or request

Status: Open.

#500 In NVIDIA/numba-cuda;

· leofang opened on Oct 2, 2025

perf: cupy arrays are slower than pytorch tensors as well as <code>DeviceNDArray</code>

bugSomething isn't workingSomething isn't working

Status: Open.

Bug

#578 In NVIDIA/numba-cuda;

· cpcloud opened on Nov 4, 2025

[FEA] Allow disabling caching in <code>compile_internal</code>

feature requestNew feature or requestNew feature or request

Status: Open.

#590 In NVIDIA/numba-cuda;

· shwina opened on Nov 14, 2025

Test: Move from unittest to fully pytest-based?

feature requestNew feature or requestNew feature or request

good first issueGood for newcomersGood for newcomers

help wantedExtra attention is neededExtra attention is needed

Status: Open.

Enhancement

#619 In NVIDIA/numba-cuda;

· leofang opened on Dec 1, 2025

[BUG] Integer overflow in device array when numpy version &gt; 2.0 for T4 GPU:s

bugSomething isn't workingSomething isn't working

Status: Open.

#623 In NVIDIA/numba-cuda;

· lcniel opened on Dec 2, 2025