Add support for compiling to multiple targets.

If a flag along the lines of "--target=sse4,avx-x2" is provided on the command-line,
then the program will be compiled for each of the given targets, with a separate
output file generated for each one.  Further, an output file with dispatch functions
that check the current system's CPU and then chooses the best available variant
is also created.

Issue #11.

run_tests.py

@@ -26,7 +26,7 @@ parser.add_option("-s", "--static-exe", dest="static_exe",
                   help="Create and run a regular executable for each test (rather than using the LLVM JIT).",
                   default=False, action="store_true")
 parser.add_option('-t', '--target', dest='target',
-                  help='Set compilation target (sse2, sse4, sse4x2, avx, avx-x2)',
+                  help='Set compilation target (sse2, sse4, sse4-x2, avx, avx-x2)',
                   default="sse4")
 parser.add_option('-a', '--arch', dest='arch',
                   help='Set architecture (x86, x86-64)',