I noticed your work on automating proofs for GPU kernels. I’ve implemented a similar concept in my acc_processor [Rust] where the vk_defcon system automatically downgrades execution if the formal invariants aren't met in the 'Thunderdome' simulation. I believe my NFE architecture could significantly accelerate the model-checking bottlenecks you’re hitting with SMT solvers."
you have funding, i live on $900 a month for 2 people
For further actions, you may consider blocking this person and/or reporting abuse
Top comments (1)
medium.com/@gaddzooks61/the-unbrib...