31–40 of 45 posts
What is MIR in this context? On the one hand, given the mention of Cranelift, it seems like it could be referring to the Rust compiler's intermediate representation, but given the context perhaps it's referring to an independent intermediate representation that also just happens to be called MIR?
Earlier quoted context omitted.
I think one of the interesting aspects of WebAssembly compared to JavaScript is that it can be efficiently AOT-compiled. I've been interested in investigating AOT compilation for a browser (perhaps there is a distant/alternative future where web browsing does not require a JIT), but maybe Wasm AOT compilers aren't really there yet.
Functionally what browsers do under the hood is AOT compilation but not in the way that i.e. Wasmtime is. The following is my knowledge that may no longer be accurate, but this is the sort of model we planned for when designing WASM to begin with: Browsers do a on-demand 'first tier' compilation for fast startup, and in the background using threads they do a high quality AOT compilation of the whole module. That high…
Wow, that dominator tree algorithm I wrote as an intern ( https://bugzilla.mozilla.org/show_bug.cgi?id=666426 ) seems to have lasted 13 years! At the time we assumed that graphs would be small enough to not warrant the constant factor against Lengauer-Tarjan. ...And of course, browser-based apps have gotten much bigger since then. Semi-NCA hadn't even been published yet and seems like the clear choice nowadays, so I'…
The compiler wasn't really designed for these huge graphs, but fortunately a lot of the issues can be fixed incrementally and it's still holding up well.
That is a stunningly large function. I've looked around the onnxruntime sources and can't find anything like it. The largest C file is under 6000 lines. Does anyone know what function it's referring to?
What is MIR in this context? On the one hand, given the mention of Cranelift, it seems like it could be referring to the Rust compiler's intermediate representation, but given the context perhaps it's referring to an independent intermediate representation that also just happens to be called MIR?
MIR is the compiler intermediate representation used by Ion. I know the IonMonkey docs say the acronym is Mid-level Intermediate Representation.
Earlier quoted context omitted.
> Is it really the case that browsers have default-enabled all sorts of extensions that are not yet widely supported by the rest of the ecosystem? I don't know the answer, but it would be hard to blame them for following normal browser development practices on the standard they created for the purpose of being in browsers.
Fair enough. I think it would be unfortunate if the WebAssembly language in browsers were a significantly different language than WebAssembly outside of browsers (just referring to language itself, not the overall runtime system). I don't think that has quite happened, and the outer ecosystem can probably catch up, but it worries me.
Now we get a couple of startups trying to make WebAssembly outside of the browser as if was a novel idea, never done before.
> control flow graph contained 132856 basic blocks That is a stunningly large function. I've looked around the onnxruntime sources and can't find anything like it. The largest C file is under 6000 lines. Does anyone know what function it's referring to?
Also the CPU emulator tick function which is a single huge switch statement with several thousand case branches was actually running into a slow path in Clang a couple of years ago (it took like 30 seconds to build that one source file). This was fixed at some point though.
I also seem to remember that Emscripten had a build setting to break up large WASM functions into smaller snippets to work around such quadratic outliers in browser WASM engines.
If anyone have a comparison with V8 that would be great!
I would love to see how Firefox compares now with the other big competitors.