[zdw]

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.

[slavik81]

The AMD Math Libraries team is hiring [1], and one of the libraries they develop is rocFFT [2]. Disclosure: I work at AMD, though not on rocFFT.

[1]: https://jobs.amd.com/job/Calgary-GPU-Libraries-Software-Deve... [2]: https://github.com/ROCmSoftwarePlatform/rocFFT

[tinus_hn]

The author lists his email address on the site and indicates he’s looking for a position.

[slavik81]

I should probably make it clear that I have nothing to do with the hiring process whatsoever. Though, it seems I could provide a referral.

[jjeaff]

Ya, but the important question is can they invert a binary tree on a whiteboard?

[umvi]

Just get a clear whiteboard, draw the binary tree, then flip the whiteboard 180 around the vertical axis so you are now looking through the back of the whiteboard.

[repsilat]

Honestly, it's an O(1) solution depending on choice of data structure, possibly constant-factor more efficient through your program depending on language, and if you're programming on a whiteboard it might also arguably be the idiomatic way to do it in that context.

Limited in that you can only have one tree per whiteboard, though.

[qppo]

"Write an FFT" is the DSP engineer interview question that's analogous to tree traversal algorithm whiteboarding. The hard part is remembering how a butterfly computation works, and you'll almost never need to implement it.

[mangamadaiyan]

... or are leetcode-proficient, these days.

[TomVDB]

One should hope that the non-CUDA GPU compute library ecosystem has already advanced beyond being able to calculate FFTs!

[singhrac]

Sure, but if Nvidia/OpenAI/Google/Facebook have shown anything, it's that there's always more kernels to invent and train bigger nets with.

[andi999]

Last time I checked there was no good fft for AMD.