triton-llvm-releases VS llvm-project

this is quickly getting better

>Much of MLIR requires compiling from source, from what I can tell

you can get apt packages from https://apt.llvm.org/ and build projects out of tree. you can also get packages from conda (https://github.com/conda-forge/mlir-feedstock). finally, if you look around on github you'll find tarred up releases too maintained by downstream users (e.g.https://github.com/ptillet/triton-llvm-releases).

you can also (as of very recently) build mlir-opt plugins just like for clang:

https://github.com/llvm/llvm-project/tree/main/mlir/examples...

The background here is that "ctpop < 2" or "ctpop == 1" (depending on zero behavior) is LLVM's canonical representation for a "power of two" check. It is used on the premise that the backend will expand it back into a cheap bitwise check and not use an actual ctpop operation. However, due to complex interactions in the backend, this does not actually happen in this case (https://github.com/llvm/llvm-project/issues/94829).

The checked project version is LLVM 18.1.0.

I think they should probably set LoopMicroOpBufferSize to a non-zero value even if its not microarchitecturally accurate. This value is used in LLVM to control whether partial and runtime loop unrolling are enabled (actually only for that). Although some targets override this default behaviour, AArch64 only overrides it to enable partial and runtime unrolling for in-order models. I've left a review comment https://github.com/llvm/llvm-project/pull/91022/files#r16026... and as I note there, the setting seems to have become very divorced from microarchitectural reality if you look at how and why different scheduling models set it in-tree (e.g. all the Neoverse cores, set it to 16 with a comment they just copied it from the A57).

In conclusion, none of the proposed changes to the Ruby version of the code makes a dent in the Crystal version. This is not entirely Crystal's doing: it uses the LLVM backend, which generates very optimized binaries.

As far as I know, libstdc++'s representation has two advantages:

First, it simplifies the implementation of `s.data()`, because you hold a pointer that invariably points to the first character of the data. The pointer-less version needs to do a branch there. Compare libstdc++ [1] to libc++ [2].

[1]: https://github.com/gcc-mirror/gcc/blob/065dddc/libstdc++-v3/...

[2]: https://github.com/llvm/llvm-project/blob/1a96179/libcxx/inc...

Basically libstdc++ is paying an extra 8 bytes of storage, and losing trivial relocatability, in exchange for one fewer branch every time you access the string's characters. I imagine that the performance impact of that extra branch is tiny, and massively confounded in practice by unrelated factors that are clearly on libc++'s side (e.g. libc++'s SSO buffer is 7 bytes bigger, despite libc++'s string object itself being smaller). But it's there.

The second advantage is that libstdc++ already did it that way, and to change it would be an ABI break; so now they're stuck with it. I mean, obviously that's not an "advantage" in the intuitive sense; but it's functionally equivalent to an advantage, in that it's a very strong technical answer to the question "Why doesn't libstdc++ just switch to doing it libc++'s way?"

This Ruby implementation is based on mruby and LLVM and it’s commercial software but cheap.

'Computer Architeture: A Quantitative Apporach" and/or more specific design types (mips, arm, etc) can be found under the Morgan Kaufmann Series in Computer Architeture and Design.

"Getting Started with LLVM Core Libraries: Get to Grips With Llvm Essentials and Use the Core Libraries to Build Advanced Tools "

"The Architecture of Open Source Applications (Volume 1) : LLVM" https://aosabook.org/en/v1/llvm.html

"Tourist Guide to LLVM source code" : https://blog.regehr.org/archives/1453

llvm home page : https://llvm.org/

llvm tutorial : https://llvm.org/docs/tutorial/

llvm reference : https://llvm.org/docs/LangRef.html

learn by examples : C source code to 'llvm' bitcode : https://stackoverflow.com/questions/9148890/how-to-make-clan...

See

https://github.com/llvm/llvm-project/issues/88344

https://fortran-lang.discourse.group/t/flang-new-how-to-forc...