2 ms·
Stratified Synthesis: Automatically Learning the x86-64 Instruction Set (2016) [pdf]
- bjornsing 4y agoExtremely cool! Have been thinking about how to go about this. It feels like a perfect test bed for automated reasoning. The high level description in the intro is very clear and makes a lot of sense.
- haskellandchill 4y agoNice I came across this in my research, also "Automatic extraction of x86 formal semantics from its natural language description" and "A Complete Formal Semantics of x86-64 User-Level Instruction Set Architecture" which summarizes some of the research. I am trying to develop a methodology for bootstrapping a bare bones IDE on any hardware with IO. Seems impossible so I'm trying to find a similar problem that is approachable.
- benreesman 4y agoThis is extremely cool. It sort of begs the question: does Intel have a golden formal model? They must right? The most recent thing I’ve read about it is The Pentium Chronicles (highly recommended incidentally), but that’s eons ago. Does anyone happen to know (and be at liberty to share) how the vendors cope with this internally?
- haskellandchill 4y agoI believe Intel has a model in the sense of model checking or symbolic execution as well as virtual simulations. Not sure what you mean by "golden model", but I don't believe they have a formal model to the level of compcert or related models in Coq/Agda/Isabelle.
- colejohnson66 4y agoWould it even be possible to formally verify such a system as a whole? From my limited experience with formal verification, even small designs can take over an hour to verify.
- tempodox 4y agoInteresting what the authors say about how complicated and inconsistent x86_64 semantics are. That specs of semantics written by humans contain errors. I can only hope that ARM has learned from Intel's mistakes and the situation is better with their CPUs. How can we hope for bug-free software when even ISA semantics aren't clear?
- als0 4y ago> I can only hope that ARM has learned from Intel's mistakes and the situation is better with their CPUs. How can we hope for bug-free software when even ISA semantics aren't clear? ARM's research division has spent a great deal of time on this topic. This is a good introduction and describes where they were at in 2018: https://alastairreid.github.io/talks/formalizing-arm-specs-ACSD-2018-06-15.pdf https://alastairreid.github.io/talks/formalizing-arm-specs-A...