Issues

• cutile: improve the autotune example

  enhancementTBD
  NVIDIA/cuda-rust#1517 · Filter by author HaoYang670 opened on Oct 14, 2026

• feat(cuda-device): add PTX cache-control streaming and eviction qualifiers for global memory

  enhancementTBD
  NVIDIA/cuda-rust#1514 · Filter by author amansahani opened on Oct 13, 2026

• bug: release-like rustflags invalidate build and test artifact reuse

  NVIDIA/cuda-rust#1508 · Filter by author ssoj13 opened on Oct 13, 2026

• bug: host-written constant globals can disappear during device LTO

  NVIDIA/cuda-rust#1507 · Filter by author ssoj13 opened on Oct 13, 2026

• Safe element casts can produce invalid bool values (Tensor::reinterpret, DeviceBuffer::cast_elem/cast_chunks)

  cuda-oxidecutile-rsruntimesafety
  NVIDIA/cuda-rust#1506 · Filter by author elibol opened on Oct 13, 2026
  

• API design: organize shared CUDA host types under semantic namespaces

  NVIDIA/cuda-rust#1505 · Filter by author elibol opened on Oct 13, 2026

• cuda-core: distinguish CUDA devices and contexts in the shared host API

  NVIDIA/cuda-rust#1504 · Filter by author elibol opened on Oct 13, 2026

• cuda-core: define boolean storage and element padding/packing policy

  cuda-oxidecutile-rsinteropruntime
  NVIDIA/cuda-rust#1503 · Filter by author elibol opened on Oct 13, 2026
  1

• Host build time grows superquadratically with kernels per #[cuda_module]

  NVIDIA/cuda-rust#1501 · Filter by author navneet83 opened on Oct 12, 2026

• feat(cuda-device): support u64 reductions and scans in cooperative_groups

  enhancementTBD
  NVIDIA/cuda-rust#1497 · Filter by author pedrozaz opened on Oct 11, 2026
  1

• [BUG]: With LTO, the build of a binary that uses a library with kernels stops with an error

  NVIDIA/cuda-rust#1495 · Filter by author fxeqxmulfx opened on Oct 11, 2026

• #[unroll] rejects a loop bound computed from a const generic (N / 8)

  NVIDIA/cuda-rust#1494 · Filter by author hhdri opened on Oct 11, 2026

• [BUG]: cuTile dimension parsing can panic instead of reporting compilation errors

  NVIDIA/cuda-rust#1485 · Filter by author uurl opened on Oct 11, 2026

• feat(cargo-oxide): select extra-rustflags by target triple

  NVIDIA/cuda-rust#1481 · Filter by author avlp12 opened on Oct 11, 2026

• [BUG]: tileiras 13.4 emits misaligned f16 loads with reshape/broadcast on sm_89

  NVIDIA/cuda-rust#1479 · Filter by author hhdri opened on Oct 10, 2026
  1

• #[unroll] over the clone budget leaves an array-indexing loop rolled and moves the array to .local, with no way to raise the budget

  NVIDIA/cuda-rust#1478 · Filter by author praveenperera opened on Oct 10, 2026

• Adding a kernel renames other kernels' shared-memory statics, so their PTX changes

  NVIDIA/cuda-rust#1477 · Filter by author praveenperera opened on Oct 10, 2026
  1

• cuTile: breakpoint on an inlined helper call slides to the next statement

  NVIDIA/cuda-rust#1476 · Filter by author elibol opened on Oct 10, 2026

• cuTile: audit generated source locations (optional follow-ups to #1379)

  NVIDIA/cuda-rust#1475 · Filter by author elibol opened on Oct 10, 2026
  1

• [BUG]: #[device] does not obey #[inline] on a generic function

  NVIDIA/cuda-rust#1471 · Filter by author fxeqxmulfx opened on Oct 10, 2026

• cuTile literal lowering changes false to true and zeroes large unsigned constants

  NVIDIA/cuda-rust#1469 · Filter by author filipkunc opened on Oct 10, 2026

• Add #[nounroll] (or #[unroll(1)]) so ptxas keeps a loop rolled

  codegencuda-oxidedevice-apisenhancementperf
  NVIDIA/cuda-rust#1468 · Filter by author praveenperera opened on Oct 9, 2026
  3

• [BUG]: FP16 m16n8k8 mma intrinsics are floored at sm_80, but Turing supports them

  bugcodegencuda-oxideintrinsics
  NVIDIA/cuda-rust#1467 · Filter by author praveenperera opened on Oct 9, 2026
  2

• Typed ld.shared / st.shared intrinsics on u32 shared addresses, so ptxas can fold immediates

  codegencuda-oxidedevice-apisenhancementintrinsicsperf
  NVIDIA/cuda-rust#1465 · Filter by author praveenperera opened on Oct 8, 2026
  2

• Ship prebuilt kernel artifacts from a library crate to users on stable Rust

  build-relatedcargo-oxidecuda-oxidedocumentationenhancementhost-apis
  NVIDIA/cuda-rust#1464 · Filter by author praveenperera opened on Oct 8, 2026
  3