Folkert de Vries

The message on an invalid target feature now suggests the most similar valid target feature names, for when it's been a while and you can't quite remember the exact target feature name.

https://github.com/rust-lang/rust/pull/153966

You can often make improvements like this by pattern matching on the surrounding code, you should give it a go when there are errors you run into that could be more helpful.

#rustlang

We're trying to change the f16 ABI on i386

https://inbox.sourceware.org/libc-alpha/DGTYN7HPA6L4.YU50VH5WZFYC@umich.edu/T/#u

Currently the ABI says to return _Float16 in a register that does not exist (on the baseline), forcing hacks on either compilers or users.

#rustlang #llvm #gcc

The cfg_select! macro is now stable, and will be part of rust 1.95.

https://github.com/rust-lang/rust/pull/149783

This is a builtin replacement for cfg_if, and is really effective at cleaning up complex configuration logic.

#rustlang

@artagnon@mathstodon.xyz that's unfortunate, our internal cleanups will have to wait for LLVM 23 then

@artagnon@mathstodon.xyz did lowering to the platform-specific instructions not make it into LLVM 22? On x86 the relevant PRs are from after the cutoff, but I thought that riscv at least made it in but can't get that to work

https://godbolt.org/z/Ercb7cMbE

Miri and const-eval can now execute c-variadic functions

https://github.com/rust-lang/rust/pull/150601

That's the last major technical hurdle for c-variadic functions. Time to stabilize this thing!

#rustlang

@harold@mastodon.gamedev.place we experimented with something like that, but the fallback is only really used during const evaluation and with the cranelift/gcc backends. So for now at least we picked the simple implementation in those cases, and trust that LLVM will normally provide something more efficient.

Niche rust feature now on nightly: carryless multiplication

https://github.com/rust-lang/rust/pull/152132

This operation is useful in some cryptography and checksum algorithms, as well as e.g. simdjson.

#rustlang #llvm

I've implemented a `simd_splat` intrinsic, a function that puts the same value in each SIMD lane. LLVM wasn't handling the previous workaround (a manual shuffle) very well.

https://github.com/rust-lang/rust/pull/151346

the intrinsic is now used by stdarch and portable_simd:

- https://github.com/rust-lang/stdarch/pull/2003
- https://github.com/rust-lang/portable-simd/pull/501

This intrinsic fixes a bunch of long-standing optimization issues, and appears to have sped up compile times for some crates too.

#rustlang