Building a web server in aarch64 assembly to give my life (a lack of) meaning

https://imtomt.github.io/ymawky/

134 46 3 months ago

imtomt.github.io

thisislife24• 2 months ago

This is an interesting project - kudos for executing it. I have to admit that when I was starting out in this field, I too fantasised about, "Would this software be faster, smaller and better in assembly?". Ofcourse, assembly programming made some sense in embedded electronics, which can be very resource constrained and even specialised for one particular application. Thinking from that aspect, perhaps you should consider making this a specialised program that runs on something like a Raspberry Pi - running such a web server directly on it, without an OS (or a very minimal OS), would make for a real cool and interesting project.

senfiaj• 2 months ago

IMHO, for servers IO (FS/DB, network, etc) is usually a greater bottleneck. Microoptimizations make sense only for CPU bound problems.

jvanderbot• 2 months ago

I am pained to think of TLS/HTTPS implemented as a hobby project in ASM, but would be impressed to see it.

thisislife24• 2 months ago

I was thinking more of the memory usage. As another project posted on HN described, they were running the web server in the RAM of an RPi ( https://news.ycombinator.com/item?id=48064312 ). Without a full-fledged OS, and a tiny (and fast) webserver coded in Assembly, more RAM would be available for the webserver and they could serve more users.

hatefulheart4• 2 months ago

What on earth are you talking about? Assembly makes sense in desktop computing as well. Have you ever, for example, watched a video? What do you think powers the codecs, JSX?

xnorswap2• 2 months ago

I'd assume most are written in C or some other higher level language. Certainly I checked the x264 ( VLC implementation of h264 ) and that is C.

thisislife24• 2 months ago

Yes, I do know that some Assembly is used in systems programming and other niches where it makes sense. To be clear, I was talking about the phase some of us go through (as amateurs) when we think everything would be "faster, smaller, better" if written in assembly - Python is slow. What about C or Pascal? Wouldn't asm be faster? ... but, as we all realise sooner or later, there's a reason we prefer to code in high-level or very high-level languages, and why pre-mature optimisation can be a real handicap.

pixelesque• 2 months ago

Previous comments from two days ago:

https://news.ycombinator.com/item?id=48080587

theanonymousone• 2 months ago

Something funny has happened. I didn't submit this link today, as is shown in the front page. I submitted three days ago, before the original author (above submission) actually.

tomku• 2 months ago

HN has a "second chance queue" that sometimes revives posts and gives them another shot. Not entirely sure how it works, but it sounds like what happened here.

globnomulous2• 2 months ago

Maybe combine the threads, dang?

peterpanhead• 2 months ago

Fun project.

DeathArrow• 2 months ago

Meanwhile another 10000 developers published desktop apps, mobile apps and system software written in javascript. /s

imvetri• 2 months ago

How to do assembly only to build a web server for the hardware? I mean it may sound like building os from scratch.

StilesCrisis• 2 months ago

"i hate string parsing. especially in assembly."

Well. Interesting choice of side project!

globnomulous2• 2 months ago

I love this. I'm wondering: what skills did this project hone? What are you better at now than you were before you undertook it? Or was it just for fun?

drob518• 2 months ago

I’m curious what the performance of this implementation is versus a server written in C, C++, or Rust. How much performance can a human still squeeze out at the assembly level versus today’s state of the art compilers?

pjdesno• 2 months ago

I’m curious what the performance of this implementation is

Almost certainly crap.

As the author states, it's a simple fork-on-request server, which was state-of-the-art in about 1996. But that's not the point.

jubilanti• 2 months ago

And that wasn't the question the parent comment asked.

mananaysiempre• 2 months ago

How much performance can a human still squeeze out at the assembly level versus today’s state of the art compilers?

Most of the squeezing is to be had in the parts where the compiler can’t help. (Which I guess is logically equivalent to saying that you can’t often do meaningfully better than the compiler on the things that the compiler is concerned with, but you have to admit it reads very differently.) Two important widely-applicable examples are data layout (locality, in particular getting rid of large and costly-to-traverse pointers) and vectorization; what they have in common is that you may well have to redesign the entire flow of data in your program around the issue before you get meaningful improvements. (And there is often an order-of-magnitude improvement to be had on a CPU-bound task, if you are willing to spend the time and effort to optimize.)

There are also specific situations where the approaches used by modern compilers work badly. The straightforward switch-based interpreter is a well-known example: modern Clang essentially turns into Clippy and goes “looks like you’re writing an interpreter, would you like me to duplicate your dispatch for you” so branch prediction works out as well as in manual assembly, but it still allocates registers a function at a time, so when the function in question is the entirety of the interpreter including the slowpaths, the regalloc sucks. Tail-call interpreters and __attribute__((cold, noinline, preserve_most)) amount to expressing the exact same control-flow graph in such a way that the compiler can digest it better, ironically by understanding less of it at any given time. This is one way that the dumb fundamental nature of the admittedly quite smart modern compiler shines through.

And in very tight loops there are still places when doing things by hand can help. For instance, when computing a histogram of byte values over a large block (for which I’m not aware of any public vectorized code that would go faster than the best scalar options) I’ve seen Clang lose as much as 20% to (contemporary) GCC on the best C implementation[1] or its straightforward manual translation to assembly, because Clang had decided it knew better which order the instructions should go in. As a less exotic case, I’ve seen GCC lose out by about 20% to (contemporary) Clang in vectorized loops because it had decided that having half the loop body be MOVs (or rather VMOVDQAs) would be a better idea than taking advantage of AVX’s ability to not overwrite either of the input arguments, and though MOVs are basically free on a superscalar they’re not that free. I’ve even seen both GCC and Clang ignore an explicit __builtin_expect() and compile a very predictable (but unavoidable) inner-loop branch into a CMOV, once again costing me about 20% in performance.

So if you do in fact care about the difference between 1.1 cycles/byte and 1.3 cycles/byte, yes you can beat a compiler even on a micro level. You just probably don’t have the, depending on your point of view, fortune or misfortune of working on code like that.

[1] https://github.com/powturbo/Turbo-Histogram

hansvm• 2 months ago

Today's state of the art compilers can't even do vectorized integer division by a compile time known constant very well. They definitely can't map high-level constructs onto low-level patterns, and they don't carry anywhere near enough semantic information through the different optimization passes to be able to take even very safe, simple, sane shortcuts with zero possibility of UB or other issues. There's a lot of performance being left on the table.

Mind you, somebody who's sympathetic to the machine's needs can easily scrape most of that performance back by writing C/C++/Zig in a way that easily maps to the optimal assembly. The optimizer won't make your code drastically worse too often, so if you start with something nice then actually dropping down into assembly has limited use cases and usually limited benefits...if you know what you're doing and throw out every style guide as you do so.

As to this server in particular? At first blush it looks more like a learning exercise. You'll go a lot further with clever incremental routines and appropriately leveraging your OS's async API than you will by shaving a few instructions here and there.

As to servers in general? Your kernel is the real bottleneck. If you need all of its features then you don't have a lot of options, but if you're like most applications then you're leaving a ton of performance on the floor not going for kernel bypass (not that using your kernel for network is a _bad_ decision, but you are nevertheless incurring a 10x-50x performance hit as the cost). Assembly shenanigans literally don't matter in comparison.

nikisweeting• 2 months ago

Nice, I've been trying to learn assembly myself recently and built a little REPL to play around with it directly https://github.com/pirate/assembly-repl

But I've struggled with the IO parts, it seems every system is so different, it's hard to live without LLVM-IR as the middle layer abstracting away compiler and target differences.

declank• 2 months ago

There is the lesser-known web server rwasa written in flat assembler: https://2ton.com.au/rwasa/

triilman• 2 months ago

I have created the same project but for Risc-V https://github.com/triilman25/tcp-socket-in-riscv-assembly

podlp• 2 months ago

The work of a true masochist, but in all seriousness, it’s a solid article explaining aarch64 assembly in plain language.

Building a web server in aarch64 assembly to give my life (a lack of) meaning

Related Stories