VkFFT – Vulkan Fast Fourier Transform Library

If I were a hiring person at AMD or Intel, I'd shortlist this guy for a job, as they need help competing against the headstart CUDA has in the GPU-base compute space.

What are the common applications for these sorts of GPU-accelerated FFTs? We mostly just solved problems analytically in undergrad, and the little bit of naive coding we did seemed pretty fast. I feel like this must be used for problems I would have learned about in grad school, if I had continued in electrical engineering.

How does using Vulkan for computation fit into the OpenCL/CUDA landscape? Is CUDA's proprietary nature doing meaningful harm, and does Vulkan help?

&quot;VkFFT aims to provide community with an open-source alternative to Nvidia&#x27;s cuFFT library, while achieving better performance.&quot;<p>There are no error bars on the graphs, so it&#x27;s very hard to judge if the minor differences are significant. I work in research, so probably I&#x27;m peculiar about this point, but: I&#x27;d expect better from anyone who&#x27;s taken basic statistics. But from a quick look, it seems like the performance is pretty much just &quot;on par&quot;.<p>It would also be nice to know how performance is on other hardware. I&#x27;m assuming it&#x27;s tuned to nvidida GPUs (or maybe even the specific GPU mentioned). But how does this perform on Intel or AMD hardware? How does it compare to `rocFFT` or Intel&#x27;s own implementation?

I think this guy will have no problem getting hired. Being conscious enough to push code online works so much better than the CV preparation courses. You know you&#x27;re on the right path when you are asked to play up your CV abstract than to downplay it.<p>Personally, I would have a hard time hiring anyone without a Github account and less so working in a place where nobody has one.

What is &quot;Native zero padding to model open systems&quot;? And how come it is &quot;up to 2x faster than simply padding input array with zeros&quot;?

Very cool!<p>Seems a bit more feature complete than my take on the problem: <a href="https:&#x2F;&#x2F;github.com&#x2F;Lichtso&#x2F;VulkanFFT" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;Lichtso&#x2F;VulkanFFT</a><p>Still, to beat CUDA with Vulkan a lot is still missing: Scan, Reduce, Sort, Aggregate, Partition, Select, Binning, etc.

Warning: LGPL license

Isn&#x27;t LGPL 2.1 is an odd license for something like this? Does it produce a library?

&gt; Support for big FFT dimension sizes. Current limits: C2C - (2^24, 2^15, 2^15),<p>What about bigger than big? &gt; 2^29 or so ? Are these sizes for double precision ?

I wonder if this works on the raspberry pi with the new Vulkan drivers.

I&#x27;m very eager to see GPU acceleration make its way into audio production, which is all still heavily CPU bound.<p>A Free GPUFFT implementation will certainly help! Great work.

may someday please someone help dethrone the underlord of AI &amp; rise us up

This guy will get a foot in but still have to do a gotcha interview loop