tbaederr wrote:

> We probably should. (LLVM gets fancy with switch lowering, but even a simple 
> binary search would solve the quadratic growth.)

Yeah, when initially landing the switch support I did a quick benchmark and it 
was slightly slower than what the current interpreter is doing. But whatever we 
do shouldn't regress the bytecode compilation time either because switch 
statements are often also small.

https://github.com/llvm/llvm-project/pull/201845
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

Reply via email to