Ingvar Stepanyan

> How does the code you are porting use bitmasks? It's somewhat complicated, but the simplified version is that it's essentially a search algorithm that needs to find first position...

> * There is a `BitsFromMask` method that returns a u64. This is available for ISAs where the SIMD width is

Come to think of it, another alternative could be to depend on `num_traits` instead, and just expose `fn to_bitmask() -> impl PrimInt`. `PrimInt` has all the required methods - https://docs.rs/num-traits/latest/num_traits/int/trait.PrimInt.html...

I think for Wasm it's fairly straightforward too - you just use i8x16.bitmask on each 128 bit vector. For Neon... Yeah, from quick search it does seem a bit more...

> and having opaque type with a generic trait Actually, I just realised there is also a place where I use pdep on the result, and pdep doesn't have a...

Hm yeah but at that point I might as well use the std::arch APIs directly, it's the abstractions that are tempting. I guess for now I can keep using wide...

Isn't that the one we talked about recently, the one you can use [Stub Libraries](https://lld.llvm.org/WebAssembly.html#stub-libraries) for?

As a quick experiment / workaround, I've used onnxscript rewriter API to insert dummy opposite transposes around each Resize node prior to saving it and manually reordered the scales: ```py...

Ah right, because of shared workspace it's not as trivial as I had hoped. Another reason to split into two workspaces for host & wasm as discussed earlier.

> but if you copy js_name and js_namespace to each item it should work Yeah, that's what I mentioned as a workaround. At the very least, it would be nice...