Improve LLVM IR Code generation

[Aatch]

This is a metabug to group together a bunch of related issues do with our current generation of LLVM IR.

The IR we produce is very verbose, this causes an inordinate amount of time to be spent during the optimization passes that we could trivially prevent by simply not outputting these sequences.

Instrumentation and testing issues

• Begin testing emitted LLVM IR #5073 : begin testing emitted LLVM
• rustc: improve -Z trans-stats to report per-fn LLVM instruction counts and translation timing #7456 : improved trans stats
• LLVM codegen-equivalence tests #7461 : LLVM codegen equivalence tests

Too many copies of functions issues

(see http://people.mozilla.org/~graydon/symbols-by-name.txt)

• rewrite type_use #8651 : rewrite type_use
• normalize_for_monomorphization could normalize recursively #2529 : normalize for monomorphization recursively
• In trans::type_use::mark_for_expr, don't count fields after the chosen field #2537 : type_use should only consider fields before the selected field
• Match checking has quadratic average complexity #7462 : match combinatorial codegen problems
• Function body is translated more than once #7349 : multiply-instantiating monomorphic closures within polymorphic functions
• Reuse default methods #8650 : reuse default methods

Too much codegen issues

• Experiment with reusing monomorphized instances across crates #1980 : cross-crate monomorphic instance reuse
• implement "dynamic drop" semantics using flags on the stack rather than zeroing #5016 : move semantics
• use tbaa.struct to inform LLVM about struct/enum padding #6736 : tbaa.struct for skipping padding
• intrinsics for trivial functions #8648 : intrinsics for trivial functions

Incorrect or LLVM-frustrating codegen issues

• Get to LLVM-lint-clean #7463 : LLVM lint-cleanness
• Fast isel dislikes switch, which pattern matching can generate #4353 : fastisel vs. switch
• Invoke instructions kick us off the FastISel path #3551 : fastisel vs. invoke
• Assembly comments affect generated code. #3298 : asm comments

Overall use of LLVM issues

• Discover better llvm optimizations for Rust #3519 : better LLVM optimizations

[graydon]

See -Z count-llvm-insns and pipe output to xdu.

[Aatch]

@graydon or, for people that do not have xdu, if you have xdiskusage instead, you can do rustc foo.rs -Z count-llvm-insns | xdiskusage

[farnoy]

Sorry for going off topic a bit, but this could improve readability of these kinds of metabugs: https://github.com/blog/1375-task-lists-in-gfm-issues-pulls-comments

[thestinger]

I think this kind of categorization is covered well by the usage of tags (I-slow, I-compiletime). It's not ever going to be possible to complete this metabug or mark it as a milestone so I don't see much value in leaving it open forever. The discussion here is out of date too.