5 ms·
I would love to be able to use Rust in my professional project. Unfortunately, I am doing high performance scientific computing. Rust doesn't even come close to
by jokteur 2y ago
I would love to be able to use Rust in my professional project. Unfortunately, I am doing high performance scientific computing. Rust doesn't even come close to offer any good alternative to cross-plateform, cross-device (CPU/GPU) libraries such as OpenMP Target, Kokkos, SYCL, ... I believe we need Nvidia/AMD to take Rust seriously (I'm not sure it is even possible without unsafe everywhere) to be able to offer good libraries.
In my world, using C++ is the modern language, because most project are stuck with Fortran.
- pdimitar 2y agoThat's completely valid and that's why I said "almost anything" -- other things include kernel development and embedded. My broader point was that C/C++ are still kings in some areas but people insist on using them where a bunch of languages do better today.
- miki123211 2y agoAlso cross-platform libraries, if you want to be everywhere, from computers and phones to set top boxes to weird Japanese rtoses you haven't even heard of, Rust just won't cut it.
- p_l 2y agoFortunately there are other languages that are more available cross-platform than Rust with its tiny Tier 1 list. Unless ofc you're contractually obligated to use vendors half-arsed kludget together patchy version of GCC of unknown provenance that they compiled with barely half C support, let alone other GCC languages.
- pjmlp 2y agoAt least C++ provides the tools to avoid this kind of problems, if one choses to do so, instead of insisting into C idioms.
- Shorel 2y agoYeah, part of the problem here is that the source code and the bug were in C, and people commenting here think C and C++ are the same thing. There's a huge difference between C and C++. There's a big difference between 2001 version of C++ and 2021 version of C++. The languages are so different that, in general, something that happens to one of them doesn't apply to the other.