how to use Numba to run z3py solver on GPU

Question

I want to run z3 solver methods such as 'solver.check()' on GPU using Numba to speed up solving. Is it possible? How can I do it?

I tested '@jit' but I can't get result.

Edit:

This is my code:

@jit
def solve(self, goal):
    solver = z3.Solver()
    solver.reset()
    solver.add(goal.as_expr())
    satisfiability = solver.check()  # Check satisfiability

goal parameter is z3 Goal object

Please provide enough code so others can better understand or reproduce the problem. — Community, Mar 27 '22 at 05:20

score 1 · Accepted Answer · answered Mar 26 '22 at 18:04

No, this is not possible. Numba does not support all Python features nor all Python libraries. Please read the documentation for more information about this. Besides this, this is also not possible to run a CPU code on the GPUs because they are completely different hardware architectures (and even if this would be possible, the resulting code would be certainly much slower on GPUs). For more information please read: this and this.

score 1 · Answer 2 · answered Mar 27 '22 at 21:13

1

Currently this is neither possible, nor it is clear exactly what one would have to do to go about to take advantage of GPUs in an SMT solver. See this discussion for further details: https://github.com/Z3Prover/z3/issues/1795

answered Mar 27 '22 at 21:13

alias

28,120
2
23
40

how to use Numba to run z3py solver on GPU

2 Answers2