5 ms·
Using ML inside compilers has a lot of untapped potential I think. People think of a compiler as an AI when they're actually very stupid in terms of the number
by mhh__ 4y ago
Using ML inside compilers has a lot of untapped potential I think.
People think of a compiler as an AI when they're actually very stupid in terms of the number of decisions available to them.
Feedback is the lifeblood of intelligent performance, it is more than possible to fake that feedback using AI.
E.g. Your error callback is on the balance of probability going to be called less than the (say) core matrix multiply loop etc. Etc.
- brrrrrm 4y agoyou might be interested in CompilerGym https://compilergym.com https://compilergym.com
- viraptor 4y agoSame for gc and database parameters. You can still get decent gains tweaking those and ml could help. I also played around with using ml to optimise auto scaling of CI instances. (taking time of day and queue sizes into account)
- titzer 4y ago> when they're actually very stupid in terms of the number of decisions available to them. I don't think this is a fair characterization, though I agree overall that ML has a lot of potential in compiler optimization.