FRESH

Hacker News

Formally Verifying Peephole Optimisations in Lean

24 points by l-mdev

by amluto

3 subcomments

Is this concept of UB as poison actually sound? It seems to me (and I apologize for not using real notation, as I can read a little bit of Lean but I’ve never tried writing it, and my experience writing optimizers is nonexistent):
Suppose we start when an SSA-style function with inputs x and y:
```
    0
```
And we rewrite it as:
```
    let z = x +nsw y --or anything else that is UB for some input
    0
```
The original code is a function that takes x and y and returns an instance of iN {n}, namely bitvec 0.
The new code also takes any input to bitvec 0, although it may discard a poison value along the way. But this is UB in the C model!
I wonder whether this situation can be avoided by declaring that there is no control flow and that the program representation disallows discarding values. (And select would need to return poison if any argument is poison even if the selector says that the poison is not used.)

by joek1301

1 subcomments

> Let’s say for moral reasons you choose to link none of LLVM’s libraries (thanks for caring about your users!)
For what moral reasons would I avoid linking LLVM? I’m not familiar.

by sapphirebreeze

0 subcomment

[dead]