Interactive GPU Programming, Part 1: Hello CUDA

From the perspective of a CUDA beginner, this doesn&#x27;t seem simpler than writing CUDA with C(not C++, just C). If you&#x27;re going to pick up CUDA, starting with C means you get the best tooling support and community docs. Not to mention that managing pointers and explicit types in C will genuinely help your understanding of how CPU-GPU works.<p>If you already know Clojure, this is probably the best chance to extend something you already love using. If you don&#x27;t, you&#x27;re probably better off learning either CUDA or Clojure rather than both at the same time. Debugging CUDA errors are already painful, I wouldn&#x27;t add a new host language on top of that.<p>For context, I&#x27;m currently taking my school&#x27;s GPGPU course. We&#x27;ve just started actually writing non-trivial code.

I was irritated that it took 8 paragraphs to reveal it was clojure. And it was extra annoying that the vagueness was entirely deliberate.

The obligatory links to the GitHub of the libraries used: <a href="https:&#x2F;&#x2F;github.com&#x2F;uncomplicate" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;uncomplicate</a>

I hope CUDA will get replaced by Vulkan merged with whatever core OpenCL features it still needs.<p>Looks like Khronos are looking into converging them in some way: <a href="https:&#x2F;&#x2F;www.pcper.com&#x2F;reviews&#x2F;Graphics-Cards&#x2F;Follow-Neil-Trevett-and-Tom-Olson-Khronos-Group-Discuss-OpenCL-and-Vulkan-Roa" rel="nofollow">https:&#x2F;&#x2F;www.pcper.com&#x2F;reviews&#x2F;Graphics-Cards&#x2F;Follow-Neil-Tre...</a><p>CUDA unfortuantely is Nvidia&#x27;s lock-in, so not a good way forward.

Followed along until I had to compile a kernel, now I&#x27;m facing a java.lang.UnsatisfiedLinkError: Error while loading native library &quot;JNvrtc-0.9.0-windows-x86_64&quot; This seems to be a dependency of ClojureCUDA, but I don&#x27;t see anything about it in their installation instructions. I have the CUDA Toolkit installed. Everything worked up to this point.

Why do things using CUDA often require older versions of compilers?<p>E.g. ccminer. Try to make it, but it finds my modern gcc or clang too modern :(

Are there similar resources&#x2F;tutorials for GPU programming&#x2F;CUDA&#x2F;openCL for pure Java?

In Polish, &quot;cuda&quot; means &quot;miracles&quot;. Always cracks me up!

I&#x27;d love something like this with Python and ctypes.

Okay, so it&#x27;s easier then directly using the the CUDA, etc., C toolchains, perhaps, but why not compare to Python + Numba, which has been available with GPU support for quite a while, and likewise avoids direct exposure to the underlying C toolchains, provides interactive compilation, can be used with a nice REPL (or, Jupyter Notebook), etc.?