sb-simd VS sleef

> fully dynamic

Well, no, it's SBCL. Common Lisp has support for types, but most compilers only use them for optimization, SBCL goes one step further and emits warnings when you mismatch types. And looking at the code, I can see lots of type declarations.

It's also interesting to note that the code does not seem to be using SBCL's new SIMD library*, so it could be sped up even more.

* <https://github.com/marcoheisig/sb-simd>, see the texinfo file for documentation.

I suppose that using arrays + using SIMD instructions could be even faster. Someone is already doing that: https://github.com/marcoheisig/sb-simd/blob/master/examples/simd-dot.lisp .

Common Lisp (sbcl) performance via native implementation of simd [0] is very impressive ! It is litteraly acheieving C/Cpp speeds (within few ms). Great work by Marco Heisig

[0] https://github.com/marcoheisig/sb-simd

Here is another demonstration of how effective SIMD vectorization can be using sb-simd.

You can see on Programming Language and Compiler Benchmark site the amazing speed of SBCL when sb-simd is used for vectorization.

For SBCL-specific, Marco and Bela have put in a ton of work at sb-simd - may be the OP finds the relevant simd interface there!

And sb-simd is getting very-very impressive to say the least thanks to Marco Heisig.

You can use sb-simd for manual vectorisation with SBCL. Manual vectorisation is definitely more hassle than automatic vectorisation, but often worth it.

Fully agreed. One more library that could open up areas is also coming soon. Though documentation is still to be written. Please check sb-simd I wish I could have supported Marco even more.

I'm the main author of Highway, so I have some opinions :D Number of operations/platforms supported are important criteria.

A hopefully unbiased commentary:

Simde allows you to take existing nonportable intrinsics and get them to run on another platform. This is useful when you have a bunch of existing code and tight deadlines. The downside is less than optimal performance - a portable abstraction can be more efficient than forcing one platform to exactly match the semantics of another. Although a ton of effort has gone into Simde, sometimes it also resorts to autovectorization which may or may not work.

Eigen and SLEEF are mostly math-focused projects that also have a portability layer. SLEEF is designed for C and thus has type suffixes which are rather verbose, see https://github.com/shibatch/sleef/blob/master/src/libm/sleef... But it offers a complete (more so than Highway's) libm.

But the other part about SIMD: I'm unsure if mgl-mat uses SIMD for transcendental functions or even for something like element-wise multiplication and division*. SIMD easily provides a speed-boost of 4-8 times which numpy uses. Libraries like sleef have been put to use by many.

Highway and Agner's VectorClass also have math functions. And SLEEF should definitely be mentioned.

"SIMD Library for Evaluating Elementary Functions, vectorized libm and DFT" - https://github.com/shibatch/sleef

Not mentioned in the list of users is SLEEF (https://github.com/shibatch/sleef), which provides fast approximations for various elementary functions. (It generates coefficients for the approximations with mpfr)

SLEEF itself is used by PyTorch.

If you are looking at floats, there's https://sleef.org

It would be interesting to see SLEEF added in the benchmarks.