Choose type for integer literals to match the target mask size (if possible).

On a target with a 16-bit mask (for example), we would choose the type
of an integer literal "1024" to be an int16.  Previously, we used an int32,
which is a worse fit and leads to less efficient code than an int16
on a 16-bit mask target.  (However, we'd still give an integer literal
1000000 the type int32, even in a 16-bit target.)

Updated the tests to still pass with 8 and 16-bit targets, given this
change.
This commit is contained in:
Matt Pharr
2013-07-23 17:01:03 -07:00
parent 9ba49eabb2
commit f7f281a256
61 changed files with 166 additions and 120 deletions

View File

@@ -37,7 +37,7 @@ parser.add_option("-g", "--generics-include", dest="include_file", help="Filenam
parser.add_option("-f", "--ispc-flags", dest="ispc_flags", help="Additional flags for ispc (-g, -O1, ...)",
default="")
parser.add_option('-t', '--target', dest='target',
help='Set compilation target (neon, sse2, sse2-x2, sse4, sse4-x2, avx, avx-x2, generic-4, generic-8, generic-16, generic-32)',
help='Set compilation target (neon, sse2, sse2-x2, sse4, sse4-x2, sse4-8, avx, avx-x2, generic-4, generic-8, generic-16, generic-32)',
default="sse4")
parser.add_option('-a', '--arch', dest='arch',
help='Set architecture (arm, x86, x86-64)',
@@ -294,7 +294,7 @@ def run_test(testname):
firstline = firstline.rstrip()
file.close()
if (output.find(firstline) == -1):
if re.search(firstline, output) == None:
sys.stderr.write("Didn't see expected error message %s from test %s.\nActual output:\n%s\n" % \
(firstline, testname, output))
return (1, 0)