Oliver is doing awesome work here. A few interesting points:
- Porffor can use typescript types to significantly improve the compilation. It's in many ways more exciting as a TS compiler.
- There's no GC yet, and likely will be a while before it gets any. But you can get very far with no GC, particularly if you are doing something like serving web requests. You can fork a process per request and throw it away each time reclaiming all memory, or have a very simple arena allocator that works at the request level. It would be incredibly performant and not have the overhead of a full GC implementation.
- many of the restrictions that people associate with JS are due to VMs being designed to run untrusted code. If you compile your trusted TS/JS to native you can do many new things, such as use traditional threads, fork, and have proper low level memory access. Separating the concept of TS/JS from the runtime is long overdue.
- using WASM as the IR (intermediate representation) is inspired. It is unlikely that many people would run something compiled with Porffor in a WASM runtime, but the portability it brings is very compelling.
This experiment from Oliver doesn't show that Porffor is ready for production, but it does validate that he is on the right track, and that the ideas he is exploring are correct. That's the imports take away. Give it 12 months and exciting things will be happing.
I'm very excited by Porffor too, but a lot of what you've said here isn't correct.
> - Porffor can use typescript types to significantly improve the compilation. It's in many ways more exciting as a TS compiler.
Proffor could use types, but TypeScript's type system is very unsound and doing so could lead to serious bugs and security vulnerabilities. I haven't kept track of what Oliver's doing here lately, but I think the best and still safe thing you could do is compile an optimistic, optimized version of functions (and maybe basic blocks) based on the declared argument types, but you'd still need a type guard to fall back to the general version when the types aren't as expected.
This isn't far from what a multi-tier JIT does, and the JIT has a lot more flexibility to generate functions for the actual observed types, not just the declared types. This can be a big help when the declared types are interfaces, but in an execution you only see specific concrete types.
> or have a very simple arena allocator that works at the request level.
This isn't viable. JS semantics mean that the request handling path can generate objects that are held from outside the request's arena. You can't free them or you'd get use-after-free problems.
> - many of the restrictions that people associate with JS are due to VMs being designed to run untrusted code
This is true to some extent, but most of the restrictions are baked into the language design. JS is a single-threaded non-shared memory language by design. The lack of threads has nothing to do with security. Other sandboxed languages, famously Java, have threads. Apple experimented with multithreaded JS and it hasn't moved forward not because of security but because it breaks JS semantics. Fork is possible in JS already, because it's a VM concept, not a language concept. Low-level memory access would completely break the memory model of JS and open up even trusted code to serious bugs and security vulnerabilities.
> It is unlikely that many people would run something compiled with Porffor in a WASM runtime
Running JS in WASM is actually the thing I'm most excited about from Porffor. There are a more and more WASM runtimes, and JS is handicapped there compared to Rust. Being able to intermix JS, Rust, and Go in a single portable, secure runtime is a killer feature.
> I haven't kept track of what Oliver's doing here lately
Please do go and check up what the state of using types to inform the compiler is (I'm not incorrect)
On the area allocator, I wasn't clear enough, as stated elsewhere this was in relation to having something similar to isolates - each having a memory space that's cleaned up on exit.
Python has almost identical semantics to JS, and has threads - there is nothing in the EMCAScript standard that would prevent them.
It is absolutely true that it is unsafe to trust TypeScript types. I've chatted briefly with Oliver on socials before and he knows this. So I am a bit confused by this issue: https://github.com/CanadaHonk/porffor/issues/234 which says "presume the types are good and have been validated by the user before compiling". This is just not a thing that's possible. Types are often wrong in subtle ways. Casts throw everything out the window.
Dart had very similar issues and constraints and they couldn't do a proper AOT compiler that considered types until they made the type system sound. TypeScript can never do that and maintain compatibility with JS.
Isolates are already available as workers. The key thing is that you can't have shared memory, other wise you can get cross-Isolate references and have all the synchronization problems of threads.
And ECMAScript is simply just specified as a single-threaded language. You break it with shared-memory threads.
In JS, this always logs '4'. With threads that's not always the case.
Based on how much imported libraries are relied upon, it makes sense to treat everything as untrusted. Unless you write every line yourself/in-house, code should be considered untrusted.
I would be curious which attack vectors change or become safe after compiling though.
The point of the js engine sandbox is to protect the user in the browser - it's completely redundant on the server. Supply chain attacks are real, but only Deno has tried to fix that through permissions/rules.
I don't think anything changes with compile to native on the server.
Totally disagree. A spec-compliant JS engine has to support the features that allow vulnerabilities like prototype pollution, which can be exploited through user input alone.
> many of the restrictions that people associate with JS are due to VMs being designed to run untrusted code. If you compile your trusted TS/JS to native you can do many new things, such as use traditional threads, fork, and have proper low level memory access. Separating the concept of TS/JS from the runtime is long overdue.
This is just outright wrong. JS limitations come from lots of things:
1. The language has almost zero undefined behavior by design. Code will essentially never behave differently on different platforms.
2. JS has traditional threads in the form of web workers. This interface exists not for untrusted code but because of thread safety. That's a language design, like channels in Go, rather than a sandboxing consideration.
3. Pretty much every non-browser JS runtime has the ability to fork.
4. JS is fully garbage collected, of course you don't get your own memory management. You can use buffers to manage your own memory if you really want to. WASM lets you manage your own memory and it can run "untrusted" code in the browser with the WASM runtime; your example just doesn't hold water. There's no way you could fiddle with the stack or heap in JS without making it not JS.
5. The language comes with thirty years of baggage, and the language spec almost never breaks backwards compatibility.
Ironically Porffor has no IO at the moment, which is present in literally every JS runtime. It really has nothing to do with untrusted code like you're suggesting.
> You can fork a process per request and throw it away each time reclaiming all memory, or have a very simple arena allocator that works at the request level. It would be incredibly performant and not have the overhead of a full GC implementation.
You also must admit that this would make Porffor incompatible with existing runtimes. Code today can modify the global state, and that state can and does persist across requests. It's a common pattern to keep in-memory caches or to lazily initialize libraries. If every request is fully isolated in the future but not now, you can end up with performance cliffs or a system where a series of requests on Node return different results than a series of requests on Porffor.
As for arena allocation, this makes it even less compatible with Node (if not intractable). If means you can't write (in JS) any code that mutates memory that was initialized during startup. If you store a reference to an object in an arena in an object initialized during startup, at the end of the request when the arena is freed you now have a pointer into uninitialized memory.
How do you tell the developer what they can and cannot mutate? You can't, because any existing variable might be a reference to memory initialized during startup. Your function might receive an object as an argument that was initialized during startup or one that's wasn't, and there's no way to know whether it's safe to mutate it.
Long story short, JS must have a garage collector to free memory, or it's not JS.
> It is unlikely that many people would run something compiled with Porffor in a WASM runtime, but the portability it brings is very compelling.
Node (via SEA in v20), bun, and deno all have built in tooling for generating a self-contained binary. Granted, the runtime needs to work for your OS and CPU, but the exact same thing could be said about a WASM runtime.
And of course there are hundreds of mature bundlers that can compile JS into a single file that runs in various runtimes without ever thinking about platform. It's weird to even consider portability of JS as a benefit because JS is already almost maximally portable.
> This experiment from Oliver doesn't show that Porffor is ready for production, but it does validate that he is on the right track, and that the ideas he is exploring are correct.
It validates that the approach to building a compiler is correct, but it says little about whether the project will eventually be usable and good. It's unlikely it'll get faster, because robust JS compatibility will require more edge cases to be handled than it currently does, and as Porffor's own README says, it's still slower than most JITted runtimes. A stable release might not yield much.
Web workers don't share memory (other than SAB) with the main thread, they are far from traditional threads. These APIs are designed the way they are to protect end users, stop sites from consuming resources or bad code blocking the main thread. None of that is needed to be that way on the server. There is zero reason that a JS implementation cannot implement proper threads within the same memory space. The issue is that all js engines are derived from the browser where that isn't wanted, they simply don't have support for it. Traditional threads need careful use,
Nowhere did I say that full, or even any, compatibility with Node is needed - it isn't.
We need to stop conflating JS the language with the runtimes.
A JS runtime absolutely can get by without a GC, you just never dealloc and consume indefinitely. That doesn't change any semantics of the language, if a value/object is inaccessible, it's inaccessible...
An arena allocator provides a route to say embedding a js-to-native app in a single threaded web server like Nginx, you don't need to share memory between what in effect become "isolates".
> Web workers don't share memory (other than SAB) with the main thread, they are far from traditional threads. These APIs are designed the way they are to protect end users, stop sites from consuming resources or bad code blocking the main thread. None of that is needed to be that way on the server.
It doesn't protect end users any more than it protects servers. Node could easily expose raw threading, but they don't because nearly the whole language isn't thread safe and everything would break. It has almost nothing to do with protecting users, it's a language design decision that enforces other design constraints.
> We need to stop conflating JS the language with the runtimes
If you're just sharing syntax but the standard library is different and essentially none of the code is compatible, it's not the same language. ECMAScript specifies all of the things you're talking about, and that is JavaScript, irrespective of the runtime.
> A JS runtime absolutely can get by without a GC, you just never dealloc and consume indefinitely. That doesn't change any semantics of the language, if a value/object is inaccessible, it's inaccessible...
If you throw away the whole heap on every request, now every request it's definitionally a "cold start". Which negates the singular benefit that this post is calling out. Porffor is still not faster than JITed engines at runtime, and initializing the code still has to happen.
> Nowhere did I say that full, or even any, compatibility with Node is needed - it isn't.
You have to square what you're saying with this statement. What you're describing is JavaScript in syntax only. You're talking about major departures from the formal language spec. Existing JavaScript code is likely to break. Why not just make a new language and call it something else, like Crystal is to Ruby? It works different, you're saying it doesn't care about compatibility... Why even call it JS then?
There is nothing in there about browser APIs, and in fact it explicitly states that the browser runtime, or any other runtime + api are not EMCAScript.
Maybe I'm not doing as much as others in my functions and I tend to stick within the AWS ecosystem, so I save some space and I presume cold-start time by not including the AWS SDK/Powertools in the output, but my functions tend to cold start and complete in ~100ms.
Sure, but the approach mentioned here benchmarks with median performance of 16ms. 100ms isn’t great especially if it’s only one part of everything that needs to happen
It’s a TS/JS to wasm to C tool chain, that runs the same JS a dozen times faster than on node. Very cool approach, and lambda cold starts are definitely where it ought to shine.
That said I wonder if it could ever go mainstream – JS is not a trivial language anymore. Matching all of its quirks to the point of being stable seems like a monstrous task. And then Node and all of its APIs are the gorilla in the room too. Even Deno had to acquiesce and replicate those with bugs and all, and it’s just based on V8 too.
Hopefully Oliver, the creator of the project, is here to better answer, but something worth bearing in mind is that a significant part of conventional JS engines are all JIT tricks. And Porffor is AOT, so can lean on WASM and the C compiler for compilation optimisations. The author did a quick comparison of lines of code for some of the popular JS engines: https://goose.icu/js-engine-sizes
We're faster! (please disregard the fact that we're barely more than a demo)
Everyone knows about 80:20, the slowdowns will come after you start doing everything your competition does.
Look at Biome. We're 15x as fast as ESLint (but disregard the fact that we don't do typeaware linting). Then comes typeaware linting and suddenly they have huge performance issues that kill the project (I'm unable to use Biome 2)
This happens over and over and over. The exceptions are very, very few (Bun is one example)
Idk, I'm having a good experience with Biome 2 in a large codebase. 4s to do a full-check including floating promises, undeclared and cyclic dependencies, and sorting imports. Our ESLint setup used to take almost a minute. The Biome team has been fixing bugs on a daily basis. Version 2.2.0 (released 3 days) ago addressed a common high-CPU-usage bug, try it out.
Edit: it's not 4s anymore, I just measured with the latest version and it takes ~900ms. Insane.
Are simply running eslint from the root of the project? And is this a monorepo?
I have used eslint in very large projects (far more than 3000 files) and running multiple instances via a task runner makes it a breeze to keep it under 30s or less, especially if you use the cache
I did try and it just stalls, taking down my CPU with it. I had moved onto other tasks and found it later still killing a few cores. It's in a monorepo with maybe 3000 files. I have no trust in the project.
Yeah I understand the frustration, I had issues setting up ESLint/Prettier years ago and spent a few hours getting the Biome configuration right. Also a monorepo with ~4000 TS/TSX files. If you ever feel like trying it again, make sure that you have "files.maxSize" and "files.ignoreUnknown" set. And be very careful with the "files.includes" list.
It sure would have been nice if biome used a better include syntax. I dislike the new way of doing it in biome 2, because it creates ambiguity and works differently than damn nearly every other tool I use
I really find it annoying when JavaScript people complain about performance. If you think something is slow then make it faster and open a pull request.
It’s great when those PRs do come, but most of the time there is just empty whining while a developer contributes nothing. This is because most JavaScript developers are deathly afraid to write original software, as that would be reinventing a wheel.
Most JavaScript developers are absolutely incapable of measuring things, so they have no idea when something else is actually faster until else runs the numbers for them. Let’s take your Bun example. Bun is great but Bun is also written in Zig which is faster than C++. Bun claims to be 3x faster at WebSockets than Nodes popular WS package, because Bun can achieve a send rate of 700,000 messages per second (numbers from 5 years ago). Bun is good at measuring things. What they don’t say is that WS is just slow. I wrote my own WebSocket library for Node in TypeScript about 5 years ago that can achieve a send rate of just under 500,000 messages per second. What they also don’t tell you is that WebSockets are 11x faster to send than to receive due to frame header interpretation. I say not to disparage Bun but to show your empty worship is misplaced if you aren’t part of the solution.
I'm not complaining about performance, I'm complaining that people say they put out something faster and smaller only to find out they're not remotely comparable.
Did you know that moment.js used to be a microlibrary? It came out replacing "the huge Date.js library."
If you've been in the field long enough, you've seen this cycle repeat over and over.
You were complaining about performance. If you, as in yourself the actual person, are not measuring things its just whining.
moment is a library discontinued years ago. They had a couple of commits in 2023 and a couple in 2022, but work stopped on the library almost 5 years ago. It is an abstraction over the big scary native Date object.
Up until 2 years ago I was writing JS full time for 15 years. I just saw a lot of people bitch and cry about inventing wheels and whining at performance targets they just guessed at. Its because most of these people could not write code. During this time there were some amazing things being written in JavaScript, and later TypeScript, but these awesome things were almost exclusively personal hobby projects the regular employed JS developer either bitched about or relied upon for career survival, because bitching and crying is what you do when you can't otherwise contribute in a meaningful way.
In my experience, the type of developer complaining about someone using JS is most of the time is some native C++/C/Rust (neck|grey)-beard. Most of the time they deflect it with "I don't support Chromium/Google" (they never elaborate why they don't contribute to Mozilla).
Lots of negativity in this thread; let me offer a bit of positivity to contrast!
The project homepage is awesome, it's a mix between a throwback to retro documentation (with ascii charts) and a console out of godbolt: https://porffor.dev/
The hangup on lack of GC is probably unnecessarily overwrought, WasmGC is pretty much here and there will be an entire ecosystem of libraries providing JS GC semantics for WASM compilers that this compiler can tap into (actually implementing the backend/runtime GC support is fairly trivial for baseline support).
It would be amazing if they pull this off. Being able to compile JS to produce minimal binaries for CLIs or just to make slim containers would be nice.
I don't want to take away from the appreciation of this awesome technical achievement, but in practice I have noticed that:
- Cold starts are kinda rare, sure it sucks that your request takes 600ms, but that means you are the first user. If you would've been served by a container that was just scaled up, you'd have been waiting for much longer
- Microservices and AWS lambda are inherently stateless, and do a ton of things to make themselves useful - get credentials, establish db connections, query configuration endpoints, all of which take time, usually more than your runtime spends booting up.
As much as I like lambdas for their deployment and operational simplicity, if you want the best UX, they have inherent technical limitations which make them the wrong choice.
I have a lambda that runs inference on a model for a feature that is seldomly used.
Lambda is far more economic given the memory requirements, but we recently moved to Rust+Candle to shave off ~300ms on cold starts as the lag could be really jarring.
It's not 200ms per call. It's 200ms cold start. Once a lambda is started, you don't pay that cost again. In most cases, you pay this a few times after a deployment or when traffic is increasing and not again.
> you pay this a few times after a deployment or when traffic is increasing and not again.
This is now how lambda works. A lambda instance sticks around for 10s of minutes at most. So you'll have cold starts every day no matter how often you deploy.
If you have very little traffic you might actually have cold starts often as the lifetime of a lambda is determined based on its use.
I've approached this historically by committing persistent/reserved instances so you always have a few instances running. This is nice on paper but feels like you're omitting what a more production-appropriate solution is. "Cold starts" aren't just slow because of init, they're also slow because that's when lots of database connections, state starts, etc happen and managing init speed won't solve that.
This is the first time that I've heard of LLRT, so here's a link for anyone else interested: https://github.com/awslabs/llrt
I still like Proffor's approach because by compiling to WASM we could have latency-optimized WASM runtimes (though I'm unsure what that might entail) that would benefit other languages as well.
Java startup times would almost certainly be worse, depending on what's going on.
A previous job I worked at ran Java on AWS Lambda. We ran our busiest Java lambda in a docker layer as our whole build system was designed around docker and from a compute performance point of view it was just as fast.
The main issues were:
* Longer init times for the JRE (including before the JIT kicks in). Our metrics had a noticeable performance hit that lined up to the startup of freshly initialized lambdas. It was still well within a tolerable range for us, though.
* Garbage collection almost never ran cleanly due the code suspension between invocations, which means we had to allocate more memory than we really needed.
The native AWS Lambda Java's 'snap start' would have helped, but the startup times were just not a big deal for our use case - we didn't bother with provisioning lambdas either. Despite the added memory costs, it was also still cheap enough that it was not really worth us investigating Java's parallel GC.
So as always, what language one should use depends on your use case. If you have an app that's sensitive to "random" delays, then you probably want something with less startup overhead.
You can use graalvm AoT. It mitigates the issues for some use cases when you need to use Java.
Standard Java on Lambda has not been feasible for me either
Rust/C++ lambda cold starts are in the 15ms ballpark, it is very unlikely that you are going to get anything much faster than that. Spinning up firecracker vm just inherently takes some time no matter what you run inside it.
I wonder if the author is aware of Node native features that improve startup times, like V8 code cache and startup snapshots. An overview of integrating them into native single-executable applications is here:
Nice we haven’t faced this cold start problem. We like the idea of Lambdas being offered in a simple runtime platform where you can store and run the code as needed.
And chain it with other stuff as well which is where workflow engines like n8n or Unmeshed.io works better. You can mix up lambdas in different languages as well.
This is exciting to see. I run some latency sensitive code on Lambda with the Node runtime, so cold starts are troublesome. I hope I'll be able to use this one it's in beta or fully released.
This is cool attempt to make a lot of JavaScript run faster in lambda. I personally got a significant decrease in cold start and runtime by switching from js to golang and would recommend as well
Use an experimental (as in, 60% of ECMA tests passing, "currently no good I/O or Node compat") AOT compiler for JS. You remove the cold start by removing the runtime, at the cost of maybe your JavaScript working and not having a garbage collector.
> It might be reasonable to go without a garbage collector if your whole lifetime is so short.
That isn't how lambda works. A single lambda instance can run for 10s of minutes handling thousands or requests. So not cleaning up memory sounds like a problem to me since you're billed by the GB/second
If you throw away your memory on every request, every request is a cold start. You've reduced your one-time startup cost but now you pay that reduced cost on every request.
No big corporate will ever use this, they'd be too worried about the compiler being compromised in some way. Llrt will never go primetime either, so we're stuck with the full Node runtime for a while.
Oliver is doing awesome work here. A few interesting points:
- Porffor can use typescript types to significantly improve the compilation. It's in many ways more exciting as a TS compiler.
- There's no GC yet, and likely will be a while before it gets any. But you can get very far with no GC, particularly if you are doing something like serving web requests. You can fork a process per request and throw it away each time reclaiming all memory, or have a very simple arena allocator that works at the request level. It would be incredibly performant and not have the overhead of a full GC implementation.
- many of the restrictions that people associate with JS are due to VMs being designed to run untrusted code. If you compile your trusted TS/JS to native you can do many new things, such as use traditional threads, fork, and have proper low level memory access. Separating the concept of TS/JS from the runtime is long overdue.
- using WASM as the IR (intermediate representation) is inspired. It is unlikely that many people would run something compiled with Porffor in a WASM runtime, but the portability it brings is very compelling.
This experiment from Oliver doesn't show that Porffor is ready for production, but it does validate that he is on the right track, and that the ideas he is exploring are correct. That's the imports take away. Give it 12 months and exciting things will be happing.
I'm very excited by Porffor too, but a lot of what you've said here isn't correct.
> - Porffor can use typescript types to significantly improve the compilation. It's in many ways more exciting as a TS compiler.
Proffor could use types, but TypeScript's type system is very unsound and doing so could lead to serious bugs and security vulnerabilities. I haven't kept track of what Oliver's doing here lately, but I think the best and still safe thing you could do is compile an optimistic, optimized version of functions (and maybe basic blocks) based on the declared argument types, but you'd still need a type guard to fall back to the general version when the types aren't as expected.
This isn't far from what a multi-tier JIT does, and the JIT has a lot more flexibility to generate functions for the actual observed types, not just the declared types. This can be a big help when the declared types are interfaces, but in an execution you only see specific concrete types.
> or have a very simple arena allocator that works at the request level.
This isn't viable. JS semantics mean that the request handling path can generate objects that are held from outside the request's arena. You can't free them or you'd get use-after-free problems.
> - many of the restrictions that people associate with JS are due to VMs being designed to run untrusted code
This is true to some extent, but most of the restrictions are baked into the language design. JS is a single-threaded non-shared memory language by design. The lack of threads has nothing to do with security. Other sandboxed languages, famously Java, have threads. Apple experimented with multithreaded JS and it hasn't moved forward not because of security but because it breaks JS semantics. Fork is possible in JS already, because it's a VM concept, not a language concept. Low-level memory access would completely break the memory model of JS and open up even trusted code to serious bugs and security vulnerabilities.
> It is unlikely that many people would run something compiled with Porffor in a WASM runtime
Running JS in WASM is actually the thing I'm most excited about from Porffor. There are a more and more WASM runtimes, and JS is handicapped there compared to Rust. Being able to intermix JS, Rust, and Go in a single portable, secure runtime is a killer feature.
> I haven't kept track of what Oliver's doing here lately
Please do go and check up what the state of using types to inform the compiler is (I'm not incorrect)
On the area allocator, I wasn't clear enough, as stated elsewhere this was in relation to having something similar to isolates - each having a memory space that's cleaned up on exit.
Python has almost identical semantics to JS, and has threads - there is nothing in the EMCAScript standard that would prevent them.
It is absolutely true that it is unsafe to trust TypeScript types. I've chatted briefly with Oliver on socials before and he knows this. So I am a bit confused by this issue: https://github.com/CanadaHonk/porffor/issues/234 which says "presume the types are good and have been validated by the user before compiling". This is just not a thing that's possible. Types are often wrong in subtle ways. Casts throw everything out the window.
Dart had very similar issues and constraints and they couldn't do a proper AOT compiler that considered types until they made the type system sound. TypeScript can never do that and maintain compatibility with JS.
Isolates are already available as workers. The key thing is that you can't have shared memory, other wise you can get cross-Isolate references and have all the synchronization problems of threads.
And ECMAScript is simply just specified as a single-threaded language. You break it with shared-memory threads.
In JS, this always logs '4'. With threads that's not always the case.
> It is absolutely true that it is unsafe to trust TypeScript types... This is just not a thing that's possible.
Well... unsafe and impossible aren't quite the same thing. I guess this is possible if you throw out "safe" as a requirement?
Based on how much imported libraries are relied upon, it makes sense to treat everything as untrusted. Unless you write every line yourself/in-house, code should be considered untrusted.
I would be curious which attack vectors change or become safe after compiling though.
The point of the js engine sandbox is to protect the user in the browser - it's completely redundant on the server. Supply chain attacks are real, but only Deno has tried to fix that through permissions/rules.
I don't think anything changes with compile to native on the server.
Totally disagree. A spec-compliant JS engine has to support the features that allow vulnerabilities like prototype pollution, which can be exploited through user input alone.
Also none of the third party code will be thread safe. Hell, some of it isn’t even reentrant.
> many of the restrictions that people associate with JS are due to VMs being designed to run untrusted code. If you compile your trusted TS/JS to native you can do many new things, such as use traditional threads, fork, and have proper low level memory access. Separating the concept of TS/JS from the runtime is long overdue.
This is just outright wrong. JS limitations come from lots of things:
1. The language has almost zero undefined behavior by design. Code will essentially never behave differently on different platforms.
2. JS has traditional threads in the form of web workers. This interface exists not for untrusted code but because of thread safety. That's a language design, like channels in Go, rather than a sandboxing consideration.
3. Pretty much every non-browser JS runtime has the ability to fork.
4. JS is fully garbage collected, of course you don't get your own memory management. You can use buffers to manage your own memory if you really want to. WASM lets you manage your own memory and it can run "untrusted" code in the browser with the WASM runtime; your example just doesn't hold water. There's no way you could fiddle with the stack or heap in JS without making it not JS.
5. The language comes with thirty years of baggage, and the language spec almost never breaks backwards compatibility.
Ironically Porffor has no IO at the moment, which is present in literally every JS runtime. It really has nothing to do with untrusted code like you're suggesting.
> You can fork a process per request and throw it away each time reclaiming all memory, or have a very simple arena allocator that works at the request level. It would be incredibly performant and not have the overhead of a full GC implementation.
You also must admit that this would make Porffor incompatible with existing runtimes. Code today can modify the global state, and that state can and does persist across requests. It's a common pattern to keep in-memory caches or to lazily initialize libraries. If every request is fully isolated in the future but not now, you can end up with performance cliffs or a system where a series of requests on Node return different results than a series of requests on Porffor.
As for arena allocation, this makes it even less compatible with Node (if not intractable). If means you can't write (in JS) any code that mutates memory that was initialized during startup. If you store a reference to an object in an arena in an object initialized during startup, at the end of the request when the arena is freed you now have a pointer into uninitialized memory.
How do you tell the developer what they can and cannot mutate? You can't, because any existing variable might be a reference to memory initialized during startup. Your function might receive an object as an argument that was initialized during startup or one that's wasn't, and there's no way to know whether it's safe to mutate it.
Long story short, JS must have a garage collector to free memory, or it's not JS.
> It is unlikely that many people would run something compiled with Porffor in a WASM runtime, but the portability it brings is very compelling.
Node (via SEA in v20), bun, and deno all have built in tooling for generating a self-contained binary. Granted, the runtime needs to work for your OS and CPU, but the exact same thing could be said about a WASM runtime.
And of course there are hundreds of mature bundlers that can compile JS into a single file that runs in various runtimes without ever thinking about platform. It's weird to even consider portability of JS as a benefit because JS is already almost maximally portable.
> This experiment from Oliver doesn't show that Porffor is ready for production, but it does validate that he is on the right track, and that the ideas he is exploring are correct.
It validates that the approach to building a compiler is correct, but it says little about whether the project will eventually be usable and good. It's unlikely it'll get faster, because robust JS compatibility will require more edge cases to be handled than it currently does, and as Porffor's own README says, it's still slower than most JITted runtimes. A stable release might not yield much.
What a strange (and strangely adversarial) comment.
Almost none of your criticisms connects with anything that the other person wrote.
> JS has traditional threads in the form of web workers.
There is no language I’m aware of where workers behave like “traditional threads”. They’re isolates. Not threads.
Web workers don't share memory (other than SAB) with the main thread, they are far from traditional threads. These APIs are designed the way they are to protect end users, stop sites from consuming resources or bad code blocking the main thread. None of that is needed to be that way on the server. There is zero reason that a JS implementation cannot implement proper threads within the same memory space. The issue is that all js engines are derived from the browser where that isn't wanted, they simply don't have support for it. Traditional threads need careful use,
Nowhere did I say that full, or even any, compatibility with Node is needed - it isn't.
We need to stop conflating JS the language with the runtimes.
A JS runtime absolutely can get by without a GC, you just never dealloc and consume indefinitely. That doesn't change any semantics of the language, if a value/object is inaccessible, it's inaccessible...
An arena allocator provides a route to say embedding a js-to-native app in a single threaded web server like Nginx, you don't need to share memory between what in effect become "isolates".
NodeJS has worker threads[0] already
[0]: https://nodejs.org/docs/latest/api/worker_threads.html
These are very similar to web workers, they don't share memory other than via SharedArrayBuffer instances. For anything else you use message passing.
> Web workers don't share memory (other than SAB) with the main thread, they are far from traditional threads. These APIs are designed the way they are to protect end users, stop sites from consuming resources or bad code blocking the main thread. None of that is needed to be that way on the server.
It doesn't protect end users any more than it protects servers. Node could easily expose raw threading, but they don't because nearly the whole language isn't thread safe and everything would break. It has almost nothing to do with protecting users, it's a language design decision that enforces other design constraints.
> We need to stop conflating JS the language with the runtimes
If you're just sharing syntax but the standard library is different and essentially none of the code is compatible, it's not the same language. ECMAScript specifies all of the things you're talking about, and that is JavaScript, irrespective of the runtime.
> A JS runtime absolutely can get by without a GC, you just never dealloc and consume indefinitely. That doesn't change any semantics of the language, if a value/object is inaccessible, it's inaccessible...
If you throw away the whole heap on every request, now every request it's definitionally a "cold start". Which negates the singular benefit that this post is calling out. Porffor is still not faster than JITed engines at runtime, and initializing the code still has to happen.
> Nowhere did I say that full, or even any, compatibility with Node is needed - it isn't.
You have to square what you're saying with this statement. What you're describing is JavaScript in syntax only. You're talking about major departures from the formal language spec. Existing JavaScript code is likely to break. Why not just make a new language and call it something else, like Crystal is to Ruby? It works different, you're saying it doesn't care about compatibility... Why even call it JS then?
> ECMAScript specifies all of the things you're talking about, and that is JavaScript, irrespective of the runtime.
I suggest you go and read the EMCAScript standard: https://ecma-international.org/publications-and-standards/st...
There is nothing in there about browser APIs, and in fact it explicitly states that the browser runtime, or any other runtime + api are not EMCAScript.
I really don't struggle that much with cold starts on Node.js/Lambda, and I don't do anything special, my build commands look like:
Maybe I'm not doing as much as others in my functions and I tend to stick within the AWS ecosystem, so I save some space and I presume cold-start time by not including the AWS SDK/Powertools in the output, but my functions tend to cold start and complete in ~100ms.Sure, but the approach mentioned here benchmarks with median performance of 16ms. 100ms isn’t great especially if it’s only one part of everything that needs to happen
It’s a TS/JS to wasm to C tool chain, that runs the same JS a dozen times faster than on node. Very cool approach, and lambda cold starts are definitely where it ought to shine.
That said I wonder if it could ever go mainstream – JS is not a trivial language anymore. Matching all of its quirks to the point of being stable seems like a monstrous task. And then Node and all of its APIs are the gorilla in the room too. Even Deno had to acquiesce and replicate those with bugs and all, and it’s just based on V8 too.
Hopefully Oliver, the creator of the project, is here to better answer, but something worth bearing in mind is that a significant part of conventional JS engines are all JIT tricks. And Porffor is AOT, so can lean on WASM and the C compiler for compilation optimisations. The author did a quick comparison of lines of code for some of the popular JS engines: https://goose.icu/js-engine-sizes
I seriously dislike this kind of comparisons.
We're faster! (please disregard the fact that we're barely more than a demo)
Everyone knows about 80:20, the slowdowns will come after you start doing everything your competition does.
Look at Biome. We're 15x as fast as ESLint (but disregard the fact that we don't do typeaware linting). Then comes typeaware linting and suddenly they have huge performance issues that kill the project (I'm unable to use Biome 2)
This happens over and over and over. The exceptions are very, very few (Bun is one example)
Idk, I'm having a good experience with Biome 2 in a large codebase. 4s to do a full-check including floating promises, undeclared and cyclic dependencies, and sorting imports. Our ESLint setup used to take almost a minute. The Biome team has been fixing bugs on a daily basis. Version 2.2.0 (released 3 days) ago addressed a common high-CPU-usage bug, try it out.
Edit: it's not 4s anymore, I just measured with the latest version and it takes ~900ms. Insane.
Are simply running eslint from the root of the project? And is this a monorepo?
I have used eslint in very large projects (far more than 3000 files) and running multiple instances via a task runner makes it a breeze to keep it under 30s or less, especially if you use the cache
I did try and it just stalls, taking down my CPU with it. I had moved onto other tasks and found it later still killing a few cores. It's in a monorepo with maybe 3000 files. I have no trust in the project.
Yeah I understand the frustration, I had issues setting up ESLint/Prettier years ago and spent a few hours getting the Biome configuration right. Also a monorepo with ~4000 TS/TSX files. If you ever feel like trying it again, make sure that you have "files.maxSize" and "files.ignoreUnknown" set. And be very careful with the "files.includes" list.
It sure would have been nice if biome used a better include syntax. I dislike the new way of doing it in biome 2, because it creates ambiguity and works differently than damn nearly every other tool I use
I really find it annoying when JavaScript people complain about performance. If you think something is slow then make it faster and open a pull request.
It’s great when those PRs do come, but most of the time there is just empty whining while a developer contributes nothing. This is because most JavaScript developers are deathly afraid to write original software, as that would be reinventing a wheel.
Most JavaScript developers are absolutely incapable of measuring things, so they have no idea when something else is actually faster until else runs the numbers for them. Let’s take your Bun example. Bun is great but Bun is also written in Zig which is faster than C++. Bun claims to be 3x faster at WebSockets than Nodes popular WS package, because Bun can achieve a send rate of 700,000 messages per second (numbers from 5 years ago). Bun is good at measuring things. What they don’t say is that WS is just slow. I wrote my own WebSocket library for Node in TypeScript about 5 years ago that can achieve a send rate of just under 500,000 messages per second. What they also don’t tell you is that WebSockets are 11x faster to send than to receive due to frame header interpretation. I say not to disparage Bun but to show your empty worship is misplaced if you aren’t part of the solution.
I'm not complaining about performance, I'm complaining that people say they put out something faster and smaller only to find out they're not remotely comparable.
Did you know that moment.js used to be a microlibrary? It came out replacing "the huge Date.js library."
If you've been in the field long enough, you've seen this cycle repeat over and over.
You were complaining about performance. If you, as in yourself the actual person, are not measuring things its just whining.
moment is a library discontinued years ago. They had a couple of commits in 2023 and a couple in 2022, but work stopped on the library almost 5 years ago. It is an abstraction over the big scary native Date object.
Up until 2 years ago I was writing JS full time for 15 years. I just saw a lot of people bitch and cry about inventing wheels and whining at performance targets they just guessed at. Its because most of these people could not write code. During this time there were some amazing things being written in JavaScript, and later TypeScript, but these awesome things were almost exclusively personal hobby projects the regular employed JS developer either bitched about or relied upon for career survival, because bitching and crying is what you do when you can't otherwise contribute in a meaningful way.
> If you think something is slow then make it faster and open a pull request.
But I want someone else to make it faster without me having to do anything.
In my experience the type of developer using JS for their Web apps rarely overlaps with someone comfortable doing high performance C++ runtime work.
In my experience, the type of developer complaining about someone using JS is most of the time is some native C++/C/Rust (neck|grey)-beard. Most of the time they deflect it with "I don't support Chromium/Google" (they never elaborate why they don't contribute to Mozilla).
Lots of negativity in this thread; let me offer a bit of positivity to contrast!
The project homepage is awesome, it's a mix between a throwback to retro documentation (with ascii charts) and a console out of godbolt: https://porffor.dev/
The hangup on lack of GC is probably unnecessarily overwrought, WasmGC is pretty much here and there will be an entire ecosystem of libraries providing JS GC semantics for WASM compilers that this compiler can tap into (actually implementing the backend/runtime GC support is fairly trivial for baseline support).
It looks similar to GraalVM for Java, right?
It would be amazing if they pull this off. Being able to compile JS to produce minimal binaries for CLIs or just to make slim containers would be nice.
I don't want to take away from the appreciation of this awesome technical achievement, but in practice I have noticed that:
- Cold starts are kinda rare, sure it sucks that your request takes 600ms, but that means you are the first user. If you would've been served by a container that was just scaled up, you'd have been waiting for much longer
- Microservices and AWS lambda are inherently stateless, and do a ton of things to make themselves useful - get credentials, establish db connections, query configuration endpoints, all of which take time, usually more than your runtime spends booting up.
As much as I like lambdas for their deployment and operational simplicity, if you want the best UX, they have inherent technical limitations which make them the wrong choice.
Awesome work, but I am genuinely curios about the use cases where the 200ms init time being a problem?
I have a lambda that runs inference on a model for a feature that is seldomly used.
Lambda is far more economic given the memory requirements, but we recently moved to Rust+Candle to shave off ~300ms on cold starts as the lag could be really jarring.
Serverless implementations where lambdas are being used for APIs so +200ms on a single call is not good.
It's not 200ms per call. It's 200ms cold start. Once a lambda is started, you don't pay that cost again. In most cases, you pay this a few times after a deployment or when traffic is increasing and not again.
> you pay this a few times after a deployment or when traffic is increasing and not again.
This is now how lambda works. A lambda instance sticks around for 10s of minutes at most. So you'll have cold starts every day no matter how often you deploy.
If you have very little traffic you might actually have cold starts often as the lifetime of a lambda is determined based on its use.
Unfortunately AWS has always been rather unspecific about this behavior but here's a link to a pluralsight blog https://www.pluralsight.com/resources/blog/cloud/how-long-do...
Maybe its a problem when you’re paying 200ms per small request and you have millions of them ?
The hypothetical 200ms is for cold starts. That means when the docker container that runs behind the lambda first starts up.
When a lambda is invocated, that docker container isn’t immediately destroyed. It’s kept around for reuse.
So there isn’t a realistic scenario where you would get a million cold starts in a month.
Some companies who actually depend on actions like this should consider fronting some cash to keep this project going.
If you have cold starts on all of your millions of requests, something is very wrong with your code.
I've approached this historically by committing persistent/reserved instances so you always have a few instances running. This is nice on paper but feels like you're omitting what a more production-appropriate solution is. "Cold starts" aren't just slow because of init, they're also slow because that's when lots of database connections, state starts, etc happen and managing init speed won't solve that.
This is the first time that I've heard of LLRT, so here's a link for anyone else interested: https://github.com/awslabs/llrt
I still like Proffor's approach because by compiling to WASM we could have latency-optimized WASM runtimes (though I'm unsure what that might entail) that would benefit other languages as well.
It'd be good if AWS Lambda provided a wasm runtime option. Cold start times for WebAssembly can be sub-millisecond.
It'd also be interesting to see comparisons to the Java and .NET runtimes on AWS Lambda.
Java startup times would almost certainly be worse, depending on what's going on.
A previous job I worked at ran Java on AWS Lambda. We ran our busiest Java lambda in a docker layer as our whole build system was designed around docker and from a compute performance point of view it was just as fast.
The main issues were:
* Longer init times for the JRE (including before the JIT kicks in). Our metrics had a noticeable performance hit that lined up to the startup of freshly initialized lambdas. It was still well within a tolerable range for us, though.
* Garbage collection almost never ran cleanly due the code suspension between invocations, which means we had to allocate more memory than we really needed.
The native AWS Lambda Java's 'snap start' would have helped, but the startup times were just not a big deal for our use case - we didn't bother with provisioning lambdas either. Despite the added memory costs, it was also still cheap enough that it was not really worth us investigating Java's parallel GC.
So as always, what language one should use depends on your use case. If you have an app that's sensitive to "random" delays, then you probably want something with less startup overhead.
You can use graalvm AoT. It mitigates the issues for some use cases when you need to use Java. Standard Java on Lambda has not been feasible for me either
Yes, We used GraalVM for data encryption project, side-effect is super fast performance. The number of req/sec also 3x.
If you want to run in a WebAssembly runtime, perhaps Wasmer can be a good choice for you.
Note: we don't support .NET or Java atm, but we support PHP and Python is about to be fully supported!
https://wasmer.io/products/edge
Rust/C++ lambda cold starts are in the 15ms ballpark, it is very unlikely that you are going to get anything much faster than that. Spinning up firecracker vm just inherently takes some time no matter what you run inside it.
https://maxday.github.io/lambda-perf/
I can't bring myself to use Rust for AWS Lambda since the runtime is still marked as "experimental".
>The Rust runtime client is an experimental package. It is subject to change and intended only for evaluation purposes.
Cloudflare offers Wasm for their Workers service. Haven't tested it though.
https://developers.cloudflare.com/workers/runtime-apis/webas...
I wonder if the author is aware of Node native features that improve startup times, like V8 code cache and startup snapshots. An overview of integrating them into native single-executable applications is here:
https://nodejs.org/api/single-executable-applications.html#s...
Nice we haven’t faced this cold start problem. We like the idea of Lambdas being offered in a simple runtime platform where you can store and run the code as needed.
And chain it with other stuff as well which is where workflow engines like n8n or Unmeshed.io works better. You can mix up lambdas in different languages as well.
This is exciting to see. I run some latency sensitive code on Lambda with the Node runtime, so cold starts are troublesome. I hope I'll be able to use this one it's in beta or fully released.
Title had me excited before I read past the first two words
Haha, love it
Paring this with Fil-C would give it a garbage collector for free?
This is cool attempt to make a lot of JavaScript run faster in lambda. I personally got a significant decrease in cold start and runtime by switching from js to golang and would recommend as well
This is seriously snappy, and impressive work.
Tl;dr
Use an experimental (as in, 60% of ECMA tests passing, "currently no good I/O or Node compat") AOT compiler for JS. You remove the cold start by removing the runtime, at the cost of maybe your JavaScript working and not having a garbage collector.
It might be reasonable to go without a garbage collector if your whole lifetime is so short. GC by decapitation. Kinda clever.
But other than that it's impossible to assess performance with such a tiny toy.
> It might be reasonable to go without a garbage collector if your whole lifetime is so short.
That isn't how lambda works. A single lambda instance can run for 10s of minutes handling thousands or requests. So not cleaning up memory sounds like a problem to me since you're billed by the GB/second
If you throw away your memory on every request, every request is a cold start. You've reduced your one-time startup cost but now you pay that reduced cost on every request.
No big corporate will ever use this, they'd be too worried about the compiler being compromised in some way. Llrt will never go primetime either, so we're stuck with the full Node runtime for a while.
Corporations use pypi and npm, I think they don't care about being compromised.