Thanks for the update on that.

So at a guess it seems that we should expect for the next few years a more 
straightforward subset of SSE4.2 support to present but not all - I 
understand of course it may be difficult to support all intrinsics in a 
cross platform way taking into account NEON differences etc.  Ideally we 
would just use our existing SSE4.2/AVX/NEON intrinsic paths for now that we 
have experience with but I suspect I may have to look into a specific path 
for this as there are some instruction oddities we rely upon with SSE etc - 
are there C language equivalents to that spec implemented already regards 
float4x4 type stuff or is it purely things like this working: 
https://gcc.gnu.org/onlinedocs/gcc/Vector-Extensions.html?

It would be good to have cross platform SIMD types and operations well 
defined by the C/C++ standards but I won't hold my breath on that one but 
it's past due - not Emscripten's problem though and in fact this all could 
help promote that happening one day.  Then I can dream of hardware support 
being made more orthogonal to suit this.

On the native 64bit support that's predictably going to be a pain as I get 
the impression we are a couple of years out from that happening then, and 
for us personally we left 32bit land behind a few years ago - but at least 
it will run in some form so thanks for that :)

The next thing would be the state of WebGL2 support but I understand that's 
very early days and really I guess we will be more interested in WebGL3 by 
the time that comes around unfortunately!

-- 
You received this message because you are subscribed to the Google Groups 
"emscripten-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to