Adventures in SIMD-Thinking - Bob Steagall - [CppNow 2021]

[@simfoo]

I'm a simple man. I see a Bob Steagall video, I click on it. Awesome presenter and you always learn a thing or two.

[@benjaminshinar9509]

this was the first time i managed to follow along a SIMD lecture. I won't be able to do it myself or explain most of it, but I didn't feel lost.

[@BobBeatski71]

Now that was an amazing tutorial on SIMD and how 'windowing' is applied to input data (signals). 👍

[@BoostCon]

Glad to hear how much you appreciated this presentation.

[@aajas]

1:22:33 - my suspicion is that clang was spilling registers, which is always something to check for. They provide an amazingly wonderful tool called llvm-mca to do analysis
llvm-mca takes as input the machine code and generates the intra-cycle pipeline analysis that it perceives a given processor architecture will execute.. (Most importantly where and why it will stall)
AVX-512 with 32 regs was always somewhat of a wild idea, so it wouldn't be surprising that clang was not optimized for it

[@hosseinaghakhani6733]

An excellent talk. Explained everything beautifully.

[@BoostCon]

Glad you liked it!

[@MaksimVolkau]

amazing galaxy brain content is here!

[@origamibulldoser1618]

This should be good :)

[@SimonToth83]

Strange results, n-th element should be O(N), so at larger inputs that should make it faster.

[@SimonToth83]

Ah, nvm, this is n-th element on 7 elements always, so yeah, the overhead will prevail then.

[@codesolution1715]

👍👍👍