alpaka

alpaka copied to clipboard

While working on #1977 (which adds device-side debug flags) I discovered that CUDA

j-stephan

[abandoned] Allowing zerocopy: `makeAvailable(queue, dst, srcView)`

5

Adds an overload set `zero_memcpy` that only copies a buffer if the destination device requires the buffer in a different memory space in order to access it. Otherwise, no copy...

bernhardmgruber

Issues Compiling with NVCC Host Compiler

5

I am trying to move my code to a newer version of Alpaka (from 0.7 to 0.9 or above) but I'm running into a compilation issue. I have code which...

GNiendorf

Plans for Half Precision / fp16 Support?

2

I was just curious if there are any plans for half precision / fp16 support in the future. Thank you.

GNiendorf

Status of atomic operations on different platforms

4

While reviewing the implementation of the atomic operations in SYCL, I started comparing what operations are available on CUDA, HIP and SYCL:  What operations should be supported by Alpaka...

fwyzard

Currently only `pow` is tested with mixed `float` and `double` arguments, in [test/unit/math/src/powMixedTypes.cpp](https://github.com/alpaka-group/alpaka/blob/develop/test/unit/math/src/powMixedTypes.cpp). All binary math functions should be tested with mixed `float`, `double` and integer arguments.

fwyzard

Consider adopting a metaprogramming library

18

During the accessor development #1249 I needed to implement a few meta functions on the side. Since alpaka is TMP heavy, we are going to need such metaprogramming facilities regularly...

bernhardmgruber

In CUDA/ROCm timing is disabled by default when [creating events](https://github.com/alpaka-group/alpaka/blob/a9f5b59da076e0371a54cb7c4158b50f116e13f5/include/alpaka/event/EventUniformCudaHipRt.hpp#L58), while in SYCL the profiling on the queue is [enabled](https://github.com/alpaka-group/alpaka/blob/a9f5b59da076e0371a54cb7c4158b50f116e13f5/include/alpaka/queue/sycl/QueueGenericSyclBase.hpp#L46). It would be nice to align all the backends to...

AuroraPerego

[RFC] move some of the CMS utilities built on top of Alpaka into Alpaka itself - or a separate library ?

6

In CMS we developed various functionality that is general enough and might be interesting to share. Some are caching solutions we developed to improve the efficiency of the original CUDA...

fwyzard

Alpaka function should accept non-const arguments by value or rvalue reference

4

Various Alpaka functions that mutate their non-`const` argument take them by non-`const` reference: - `ALPAKA_FN_HOST static auto createTaskMemset(TView& view, std::uint8_t const& byte, TExtent const& extent)` takes `view` by non-`const` reference...

fwyzard