27 Apr
2012
27 Apr
'12
8:29 p.m.
On 4/26/12 10:22 AM, Ryan Newton wrote:
It doesn't seem like enough of a code size reduction to justify the change in this case.
Is there any opportunity to attack this problem later in the compiler? Perhaps a CSE for similar blocks of code? I've noticed enormous reductions in size using UPX, so I know these binaries themselves are quite compressible.
I agree. The reduction doesn't seem large enough to justify a change. Though it definitely seems like we should keep an eye out for ways of doing CSE, or similar, in order to remove redundancies in inlining. -- Live well, ~wren