LLVM Clang 12 Performance Benchmarks On AMD Ryzen 9 5950X

[phoronix]

Phoronix: LLVM Clang 12 Performance Benchmarks On AMD Ryzen 9 5950X

With LLVM 12 due for release next month and GCC 11 not being far behind, it's the season for fresh compiler benchmarks. In today's article is a look at the AMD Ryzen 9 5950X (Zen 3) compiler performance between LLVM Clang 11.1 against the current LLVM Clang 12.0 Git development code in its feature-frozen state.

http://www.phoronix.com/vr.php?view=29958

[kvuj]

Ever since I saw that presentation from John Hennessy (https://youtu.be/Azt8Nc-mtKM?t=647) where he compared native C code using AVX to do matrix multiply to python and it was 62 806 times faster, it changed my view on computing.

I seriously hope we start using the newer and specialized instructions in our CPUs for everyday use. glibc's hw-caps might be good first step, but I'm starting to think the gentoo guys aren't as crazy as I originally thought.

[ejr2122]

One aspect of Ryzen CPUs that appeals to me is their support for ECC, which some motherboards facilitate. This makes them a good candidate for small/fast ZFS applications. Unfortunately, most ECC RAM is significantly slower than what Ryzen CPU benchmarks typically use. I would really like to know what these numbers would be like with 2666 MHz ECC RAM, and also how that RAM speed bottleneck looks with other Ryzen CPUs like the 5900X, 3950X, 3900X, and 3900XT. The bottleneck might affect whether or not the 5950X is the best choice for this use-case.