Porting a Ray Tracer to Rust, Part 1

Re:Porting a Ray Tracer to Rust, Part 1 by yuriks on Wednesday December 31, 2014 @07:23AM (#8819076)

I also did a similar thing with SmallVCM, another "educational" ray-tracer: https://github.com/yuriks/SmallVCM-rs

There is also a written report with my experience and benchmark results, but unfortunately it's in Portuguese: https://github.com/yuriks/tg-yuriks-2014b/blob/master/final/...

One thing that surprised me a lot is that Rust ended up being noticeably faster than C++ (around 20%) even when compared with clang.

[ Parent | on HN ]

Re:Porting a Ray Tracer to Rust, Part 1 by berkut on Wednesday December 31, 2014 @07:32AM (#8819094)

Do you mean compared to C++ compiled with LLVM (as opposed to GCC)?

I've found that LLVM is generally faster than GCC and in some cases significantly so.

[ Parent | on HN ]

Re:Porting a Ray Tracer to Rust, Part 1 by yuriks on Wednesday December 31, 2014 @07:44AM (#8819122)

Yes, I compared Rust and Clang using similar LLVM versions and the Rust version of the code consistently beat Clang.

[ Parent | on HN ]

Re:Porting a Ray Tracer to Rust, Part 1 by darkpore on Wednesday December 31, 2014 @08:00AM (#8819170)

Not the same version then? :)

[ Parent | on HN ]

Re:Porting a Ray Tracer to Rust, Part 1 by kibwen on Wednesday December 31, 2014 @01:12PM (#8820370)

For the moment Rust is using a custom fork of LLVM while waiting for its patches to be upstreamed.

[ Parent | on HN ]

Re:Porting a Ray Tracer to Rust, Part 1 by zxcdw on Wednesday December 31, 2014 @02:50PM (#8820797)

Hasn't this been the case like forever (at least for the past two years)?

[ Parent | on HN ]

Re:Porting a Ray Tracer to Rust, Part 1 by kibwen on Thursday January 01, 2015 @09:07PM (#8825195)

They keep coming up with new patches to upstream. :)

[ Parent | on HN ]