[LinalgExt] Argsort leads to dynamic-sized stack allocation

Crash in `iree-reduce` with complex tensors

Masked vectorization of a contiguous `tensor.extract` produces an invalid `vector.transfer_read` mask when the source rank < loop rank

[Bug] Pool ceil_mode=1 causes tensor.cast shape mismatch compilation failure (AveragePool, MaxPool, 1D/2D/3D)

ReduceMax/ReduceMin over-propagates NaN — returns NaN for rows where non-NaN value should win

Sign operator returns 0 for NaN inputs instead of NaN

[AMDGPU/RDNA4] End-to-end support for global_load_tr (global memory transpose load)

resnet50-v1-12 / resnet50-v2-7 produce all-NaN output on hip/gfx1100 when batch dim is dynamic

IREE produces NaNs in onnx.Det test because of how torch-mlir lower the op

Inconsistent DMA benefit between --iree-rocm-target=gfx950 and --iree-rocm-target=mi355x