Rust programs versus Go • Hacker News Board

Re: Rust programs versus Go

#21

Post by v_lisivka » Wed, May 16, 2018, 9:57 AM UTC

weavie wrote: post #16

simias wrote: post #8

Earlier quoted context omitted.

It's true that performance is not always the main metric to consider but in this case I'm not really sure that Go makes that much of a difference. Take the "regex-redux" benchmark for instance, compare the two versions: https://benchmarksgame-team.pages.debian.net/benchmarksgame/... https://benchmarksgame-team.pages.debian.net/benchmarksgame/... I don't find the code complexity obviously worse for the Rust version, y…

You aren't really going to experience Rusts complexity when your program is essentially just a single function, so it probably isn't really a fair comparison..

When programs grows in complexity, Rust shines. Extra guarantees are very handy when code refactored. It hard to compile Rust code, but when it compiled, it usually works (or quickly points out to problem).

I actually use it in parallel with C/C++ legacy code, and I just trying to avoid C/C++ as much as I can, because I newer saw segfault in safe Rust code so far.

Re: Rust programs versus Go

#22

Post by tex0 » Wed, May 16, 2018, 9:58 AM UTC

weavie wrote: post #13

simias wrote: post #8

Earlier quoted context omitted.

It's true that performance is not always the main metric to consider but in this case I'm not really sure that Go makes that much of a difference. Take the "regex-redux" benchmark for instance, compare the two versions: https://benchmarksgame-team.pages.debian.net/benchmarksgame/... https://benchmarksgame-team.pages.debian.net/benchmarksgame/... I don't find the code complexity obviously worse for the Rust version, y…

Go seems to perform really poorly in the regex-redux. Against both Node and Python Go is way faster (as one would expect) apart from with the regex-redux.. https://benchmarksgame-team.pages.debian.net/benchmarksgame/... https://benchmarksgame-team.pages.debian.net/benchmarksgame/... Does Go just suck with it's regex performance? Would this be easy to fix by just pulling in a C regex library?

The benchmark is using some third party PCRE lib. This is probably necessary because Go's RE2 lacks some features, but it's not a good comparison.

Re: Rust programs versus Go

#23

Post by smcl » Wed, May 16, 2018, 9:59 AM UTC

weavie wrote: post #16

simias wrote: post #8

Earlier quoted context omitted.

It's true that performance is not always the main metric to consider but in this case I'm not really sure that Go makes that much of a difference. Take the "regex-redux" benchmark for instance, compare the two versions: https://benchmarksgame-team.pages.debian.net/benchmarksgame/... https://benchmarksgame-team.pages.debian.net/benchmarksgame/... I don't find the code complexity obviously worse for the Rust version, y…

You aren't really going to experience Rusts complexity when your program is essentially just a single function, so it probably isn't really a fair comparison..

Especially when you're just reading some pre-written code. I've rarely (never?) seen any complaints about how reading Rust is hard ... but lots which say that writing the code itself can be frustrating

Re: Rust programs versus Go

#24

Post by ysleepy » Wed, May 16, 2018, 10:00 AM UTC

Aissen wrote: post #20

Anyone has tried modifying the go regex benchmark to use https://github.com/BurntSushi/rure-go ? It seems to use a library that isn't even on github anymore.

"Go bindings to Rust's regex engine." - I't would not be in the spirit of the comparison. - Even though most scripting languages also have native regex implementations.

Re: Rust programs versus Go

#25

Post by boyter » Wed, May 16, 2018, 10:00 AM UTC

Not really that surprising. I spent a good chunk of time trying to make a Go program as fast as a well written Rust one (tokei) which I wrote about here https://boyter.org/posts/sloc-cloc-code/ and in the end the GC was what held me back.

I suspect its possible to get Go programs for the most part close to the performance of Rust but in the end the lower level and fewer abstractions mean a well written Rust or C/C++ program will almost always be faster.

I wonder how different these benchmarks would be if the Go runtime was set to have GC disabled though. A fairly simple thing to try out and see.

Re: Rust programs versus Go

#26

Post by Dowwie » Wed, May 16, 2018, 10:05 AM UTC

rqs wrote: post #19

Dowwie wrote: post #18

And that's not all! We really need to see memory consumption.

You knew what does the word "mem" on that page means do you :D

What mem?

Re: Rust programs versus Go

#27

Post by ysleepy » Wed, May 16, 2018, 10:06 AM UTC

I would be interested in results of go-llvm - measuring some proxy of the intrinsic overhead due to the design choices of go vs. rusts focus on zero-cost abstractions.

The clean-room compiler backend of go has some usability advantages, but sacrifices the sort of comprehensive optimization that llvm does.

Re: Rust programs versus Go

#28

Post by logicallee » Wed, May 16, 2018, 10:07 AM UTC

cmyr wrote: post #10

Nothing in these results should really surprise anyone. Go is a good tool for writing networking services, but it is not (and does not intend to be) a systems language. For another perspective, look at Go versus Python: https://benchmarksgame-team.pages.debian.net/benchmarksgame/...

to be fair, it was explicitly designed to be a systems language, and its specification says "Go is a general-purpose language designed with systems programming in mind."

This should be removed, and I would say that it gave up on its design goal of being a systems language.

Re: Rust programs versus Go

#29

Post by codetrotter » Wed, May 16, 2018, 10:11 AM UTC

slezyr wrote: post #11

simias wrote: post #3

I expect Rust to be faster than Go as a rule because of its somewhat lower level nature and cost-less "unmanaged" abstractions but I didn't expect it to be a full order of magnitude faster than Go for some benchmarks. Are the Go snippets not properly optimized or is there something else going on? I think pitting Go against Rust is somewhat inflammatory but if you look at the Rust vs. C++ benchmark the results are muc…

> Rust vs. C++ benchmark the results are much more similar Not really... Try compare sources of those programs. For example spectral-norm uses OpenMP and other things that make source look really different from the one you would write in real life. Rust program looks much simpler and closer to real life code. https://benchmarksgame-team.pages.debian.net/benchmarksgame/... https://benchmarksgame-team.pages.debian.net/…

> > the results are much more similar

Parent is talking about performance, not source code

Re: Rust programs versus Go

#30

Post by codetrotter » Wed, May 16, 2018, 10:14 AM UTC

Dowwie wrote: post #26

rqs wrote: post #19

Earlier quoted context omitted.

You knew what does the word "mem" on that page means do you :D

What mem?

In portrait mode on my mobile device that website shows only the amount of time. In landscape it shows more columns of data. That kind of design is quite irritating and not how to do responsive websites right.

Perhaps you were looking at the page in portrait mode on a mobile device as well?