for parallel builds so that we space out the large links.

llvm-svn: 34098

llvm-svn: 34097

llvm-svn: 34095

mention llvm-ld and opt instead (if appropriate).

llvm-svn: 34094

llvm-svn: 34093

llvm-svn: 34092

llvm-svn: 34091

of a shell script to stand in its place. The shell script just pawns the
execution off to llvm-ld.

llvm-svn: 34090

doesn't get mingled with the output bytecode.

llvm-svn: 34089

llvm-svn: 34088

doesn't get mingled with the output bytecode.

llvm-svn: 34087

llvm-svn: 34086

llvm-svn: 34085

2. All function-level constants are now incorporated into the module-level
  constant pool, since the compaction table was removed.  Eliminate extra
  work to check for them.

This speeds up the bcwriter from 24.4s to 13.1s on 447.dealII and .73 -> .56s
on kc++ in a release build.

llvm-svn: 34084

methods as those global function's internal implementation.

llvm-svn: 34083

llvm-svn: 34082

llvm-svn: 34081

llvm-svn: 34080

llvm-svn: 34079

llvm-svn: 34076

llvm-svn: 34075

llvm-svn: 34074

llvm-svn: 34073

llvm-svn: 34072

llvm-svn: 34067

llvm-svn: 34066

llvm-svn: 34065

strings.

llvm-svn: 34064

llvm-svn: 34063

llvm-svn: 34062

Copy a couple more missing options from gccld to llvm-ld.

llvm-svn: 34061

Make llvm-ld more gccld-like by having it run the same set of passes. The
delta was probably due to lack of llvm-ld being maintained. Just another
reason to have only one optimizing linker in in LLVM.

llvm-svn: 34058

llvm-svn: 34056

llvm-svn: 34055

llvm-svn: 34054

add some signed/unsigned arithmetic operation functions into APInt.h to
handle the signed/unsigned issue. These functions will be defined inside a
namespace "APIntOps" which is inside llvm namespace.

llvm-svn: 34053

llvm-svn: 34052

llvm-svn: 34051

llvm-svn: 34050

llvm-svn: 34049