Issues

• Document API and backend synchronization semantics

  UWHPC/xpu#44 · Filter by author AlyanSal opened on Sep 30, 2026

• Default XPU_ARCH_NATIVE off for consumers

  UWHPC/xpu#43 · Filter by author AlyanSal opened on Sep 30, 2026

• Add releases, version tags and a changelog

  UWHPC/xpu#42 · Filter by author AlyanSal opened on Sep 30, 2026

• Install a usable CMake package configuration

  UWHPC/xpu#41 · Filter by author AlyanSal opened on Sep 30, 2026

• Add recoverable errors for allocation, launch and linear algebra

  UWHPC/xpu#40 · Filter by author AlyanSal opened on Sep 30, 2026

• Expose occupancy and launch-size validation utilities

  UWHPC/xpu#39 · Filter by author AlyanSal opened on Sep 30, 2026

• Seed a consecutive range of random streams on device

  UWHPC/xpu#38 · Filter by author AlyanSal opened on Sep 30, 2026

• Add segmented and general transform reductions

  UWHPC/xpu#36 · Filter by author AlyanSal opened on Sep 30, 2026

• Export CUDA extended-lambda support and a portable lambda macro

  UWHPC/xpu#37 · Filter by author AlyanSal opened on Sep 30, 2026

• Add host mirrors, pinned staging and queue-aware copies

  UWHPC/xpu#35 · Filter by author AlyanSal opened on Sep 30, 2026

• Introduce streams/queues and explicit synchronization

  UWHPC/xpu#34 · Filter by author AlyanSal opened on Sep 30, 2026

• Replace or avoid the CUDA square transpose kernel

  UWHPC/xpu#32 · Filter by author AlyanSal opened on Sep 30, 2026

• Add portable team execution and team reductions

  UWHPC/xpu#33 · Filter by author AlyanSal opened on Sep 30, 2026

• Add batched CUDA LU and inversion with reusable workspace

  UWHPC/xpu#31 · Filter by author AlyanSal opened on Sep 30, 2026

• Keep CUDA LU status on device and reuse solver handles

  UWHPC/xpu#30 · Filter by author AlyanSal opened on Sep 30, 2026

• Prefer an optimized BLAS/LAPACK implementation in CMake

  UWHPC/xpu#29 · Filter by author AlyanSal opened on Sep 30, 2026

• Specialize CUDA uniform_index for 32-bit bounds

  UWHPC/xpu#27 · Filter by author AlyanSal opened on Sep 30, 2026

• Remove row-major LAPACKE translation and NaN scans from CPU LU

  UWHPC/xpu#28 · Filter by author AlyanSal opened on Sep 30, 2026

• Support pooled or stream-ordered CUDA allocation

  UWHPC/xpu#26 · Filter by author AlyanSal opened on Sep 30, 2026

• Add uninitialized buffers for overwrite-only storage

  UWHPC/xpu#25 · Filter by author AlyanSal opened on Sep 30, 2026

• Avoid software 64-bit division in multidimensional CUDA ranges

  UWHPC/xpu#24 · Filter by author AlyanSal opened on Sep 30, 2026

• Adapt CPU reduction grain to available threads

  UWHPC/xpu#23 · Filter by author AlyanSal opened on Sep 30, 2026

• Add serial and nested-region fast paths for CPU launches

  UWHPC/xpu#22 · Filter by author AlyanSal opened on Sep 30, 2026

• Cache SoA view stride and remove checks from device accessors

  UWHPC/xpu#21 · Filter by author AlyanSal opened on Sep 30, 2026

• Offer opt-in vectorized transcendental math

  UWHPC/xpu#20 · Filter by author AlyanSal opened on Sep 30, 2026