Simple functions are way bigger than they need to be. Fix that.

[GoogleCodeExporter]

We currently know of a couple inefficiencies:

1. Because we always load the stack pointer, LLVM can't optimize away the
unwind loop, even in simple functions that, say, return once from the outer
scope. For non-generators, we can do better.
2. The block stack is stored in the frame, which again prevents LLVM from
optimizing away most of the unwind block.

This issue will hold a summary of this kind of issue, but we may split it
up when we get around to optimizing these things.

Original issue reported on code.google.com by jyass...@gmail.com on 21 Apr 2009 at 10:01

[GoogleCodeExporter]

Original comment by collinw on 27 May 2009 at 9:44

• Added labels: Performance

[GoogleCodeExporter]

Original comment by collinw on 27 May 2009 at 11:08

• Added labels: Type-Defect

[GoogleCodeExporter]

As of r655, LLVM's analyzers now have access to the block stack.

With this patch, the empty function shrinks from 168 lines of LLVM assembly to 
118, 
and 'def foo(a): return 1+a' shrinks from 294 to 201.

Original comment by collinw on 18 Jun 2009 at 8:28

• Added labels: Priority-Medium

[GoogleCodeExporter]

r663 makes the AST->bytecode compiler omit 
SETUP_LOOP/POP_BLOCK/BREAK_LOOP instructions in some cases when they're not 
needed. This is currently a function-wide analysis, and could/should be made 
more 
powerful to analyze individual loops.

For a simple function

def foo():
    for x in range(3):
        bar(x)

this patch reduces the number of non-blank LLVM IR lines from 508 to 350.

Original comment by collinw on 21 Jun 2009 at 2:49