Nick Lewycky
391b3e4b4c
Make this explictly signed. Fixes PR1571.
...
llvm-svn: 40569
2007-07-28 16:43:10 +00:00
Dan Gohman
0252aa07ee
Re-apply 40504, but with a fix for the segfault it caused in oggenc:
...
Make the alignedload and alignedstore patterns always require 16-byte
alignment. This way when they are used in the "Fs" instructions, in which
a vector instruction is used for a scalar purpose, they can still require
the full vector alignment. And add a regression test for this.
llvm-svn: 40555
2007-07-27 17:16:43 +00:00
Evan Cheng
cb8f08ebca
Reverting 40504 for now. It's breaking oggenc.
...
llvm-svn: 40547
2007-07-27 01:37:47 +00:00
Owen Anderson
ed3e63a203
Fix a bug introduced in my last commit.
...
llvm-svn: 40542
2007-07-26 18:57:04 +00:00
Owen Anderson
f2b10d3de3
Fix a couple more bugs in the phi construction by pulling in code that does
...
almost the same things from LCSSA.
llvm-svn: 40540
2007-07-26 18:26:51 +00:00
Evan Cheng
7b20c11ccf
Test case for PR1573.
...
llvm-svn: 40539
2007-07-26 17:45:57 +00:00
Evan Cheng
a681654ef4
Fix test.
...
llvm-svn: 40536
2007-07-26 17:07:03 +00:00
Dan Gohman
513dcba4f8
Remove X86ISD::LOAD_PACK and X86ISD::LOAD_UA and associated code from the
...
x86 target, replacing them with the new alignment attributes on memory
references.
llvm-svn: 40504
2007-07-26 00:31:09 +00:00
Owen Anderson
ed75133924
Fix what is _hopefully_ the last corner case for loops.
...
llvm-svn: 40503
2007-07-25 23:54:42 +00:00
Owen Anderson
1e53615073
My last commit was not correct for nested loops. Fix it, and add a testcase for it.
...
llvm-svn: 40498
2007-07-25 22:19:40 +00:00
Owen Anderson
ce9d73dbe7
Fix an infinite loop on 300.twolf.
...
llvm-svn: 40497
2007-07-25 22:03:06 +00:00
Owen Anderson
b215074820
Fix a bug in non-local memdep that was causing an infinite loop on 175.vpr.
...
llvm-svn: 40495
2007-07-25 21:26:36 +00:00
Owen Anderson
522f7a7608
Fix a bug that was causing GVN to crash on 252.eon.
...
llvm-svn: 40494
2007-07-25 21:13:41 +00:00
Owen Anderson
6a1a8d05b8
Add basic support for performing whole-function RLE.
...
Note: This has not yet been thoroughly tested. Use at your own risk.
llvm-svn: 40489
2007-07-25 19:57:03 +00:00
Chandler Carruth
581a82e0bb
Moving regression test to reflect move in source and headers to Bitcode.
...
llvm-svn: 40488
2007-07-25 19:47:31 +00:00
Dan Gohman
a2e07a38bc
Use movaps to load a v4f32 build_vector of all-constant values into a
...
register instead of loading each element individually.
llvm-svn: 40478
2007-07-24 22:55:08 +00:00
Owen Anderson
eb9f1b612c
Add a GVN pass, using the value numbering code I developed for GVNPRE and the
...
load elimination code from RedundantLoadElimination.
llvm-svn: 40469
2007-07-24 17:55:58 +00:00
Devang Patel
aa2a753350
Unreachable block is not a root node in post dominator tree.
...
llvm-svn: 40458
2007-07-24 01:02:25 +00:00
Owen Anderson
5969a3cb91
Rename a lot of things to change FastDLE to RedundantLoadElimination.
...
llvm-svn: 40457
2007-07-24 00:17:04 +00:00
Chandler Carruth
f4eead2a50
Fix for PR1567, which involves a weird bug on non-32bit architectures and silly C type sizes.
...
llvm-svn: 40451
2007-07-23 22:42:15 +00:00
Owen Anderson
004db891dd
Add testcases for FastDLE.
...
llvm-svn: 40449
2007-07-23 22:18:05 +00:00
Owen Anderson
c133ed619f
Move these tests to use FastDSE instead of old DSE.
...
llvm-svn: 40444
2007-07-23 20:49:13 +00:00
Dan Gohman
e2dcfd1516
Delete the svn:executable property on these files, which aren't executable.
...
llvm-svn: 40441
2007-07-23 19:26:08 +00:00
Duncan Sands
5a29993695
The Ada f-e produces various auxiliary output files
...
that cannot be suppressed and cannot be redirected:
they are dumped in the current working directory.
When running the testsuite this means that these
files do not end up in the Output directory. The
best solution I could find is to change directory
into Output before running tests.
llvm-svn: 40437
2007-07-23 15:23:35 +00:00
Duncan Sands
ec44ba07b5
For multipart tests, place the parts with no
...
RUN line in Support. Give up on sending output
to /dev/null - this cannot always be arranged.
llvm-svn: 40436
2007-07-23 13:41:53 +00:00
Duncan Sands
644d0388fc
Remove -O1 in favour of making llvm-gcc-4.2 a bit
...
smarter.
llvm-svn: 40391
2007-07-22 10:25:44 +00:00
Chris Lattner
b4bf211c23
This xform isn't safe, removing it.
...
llvm-svn: 40378
2007-07-21 21:27:27 +00:00
Duncan Sands
925762f19b
Need -O1 or better to have these builtins lowered to
...
llvm intrinsics in llvm-gcc-4.2. This is because
get_pointer_alignment bails out: it relies on TER
to compute accurate alignment information.
llvm-svn: 40128
2007-07-20 22:39:06 +00:00
Dan Gohman
87107326f6
Optimize alignment of loads and stores.
...
llvm-svn: 40102
2007-07-20 16:34:21 +00:00
Dan Gohman
df9edb66c2
Update these regression tests to accomodate X86InstrSSE.td now using movups/movaps
...
for everything.
llvm-svn: 40101
2007-07-20 16:31:26 +00:00
Evan Cheng
d846db5ab8
New test.
...
llvm-svn: 40077
2007-07-20 00:27:56 +00:00
Evan Cheng
d7a079b7e2
New test.
...
llvm-svn: 40073
2007-07-19 23:53:50 +00:00
Evan Cheng
b95db5ea4b
Try fixing it again.
...
llvm-svn: 40072
2007-07-19 23:53:29 +00:00
Reid Spencer
44d34e9b61
For PR1553:
...
Change the keywords for the zext and sext parameter attributes to be
zeroext and signext so they don't conflict with the keywords for the
instructions of the same name. This gets around the ambiguity.
llvm-svn: 40069
2007-07-19 23:13:04 +00:00
Bill Wendling
649bfb8c03
Don't need the "&&" to glue lines together.
...
llvm-svn: 40063
2007-07-19 18:06:26 +00:00
Duncan Sands
f5054e4d83
Fix this test. We produce pretty cruddy code for this
...
nowadays, unfortunately it's not so easy to get good
code here while still producing correct code in exotic
situations.
llvm-svn: 40050
2007-07-19 09:38:34 +00:00
Bill Wendling
ca081ee3f7
Testcase for PR1549
...
llvm-svn: 40041
2007-07-19 06:31:11 +00:00
Devang Patel
988a83909e
New test.
...
llvm-svn: 40023
2007-07-18 23:47:02 +00:00
Evan Cheng
626af8e187
New test.
...
llvm-svn: 40020
2007-07-18 21:39:16 +00:00
Dan Gohman
2fc9d5171e
Implement initial memory alignment awareness for SSE instructions. Vector loads
...
and stores that have a specified alignment of less than 16 bytes now use
instructions that support misaligned memory references.
llvm-svn: 40015
2007-07-18 20:23:34 +00:00
Dan Gohman
4c140b7128
It's not necessary to do rounding for alloca operations when the requested
...
alignment is equal to the stack alignment.
llvm-svn: 40004
2007-07-18 16:29:46 +00:00
Reid Spencer
beabff6111
Fix this test to not rely on the path but to use the
...
configured llvm-gcc instead.
llvm-svn: 39992
2007-07-18 04:37:24 +00:00
Evan Cheng
0c49836a9f
Fix test.
...
llvm-svn: 39976
2007-07-17 18:16:09 +00:00
Tanya Lattner
1b070ae252
Don't use * in XFAIL line unless you want it to XFAIL on everything. I don't believe you can use true regular expressions here. It will just look for those keywords. Please test this on sparc and see if they still are xfailed.
...
llvm-svn: 39975
2007-07-17 18:08:32 +00:00
Gabor Greif
5bf00849d3
XFAIL these on sparc-solaris
...
llvm-svn: 39969
2007-07-17 10:25:31 +00:00
Evan Cheng
532b73311e
Use push / pop for prologues and epilogues.
...
llvm-svn: 39967
2007-07-17 07:59:08 +00:00
Reid Spencer
175fd0a317
For PR1558:
...
Move tests that have C/C++ sources into the appropriate directory. This
allows them to be selected for testing based on whether llvm-gcc is
present or not.
llvm-svn: 39963
2007-07-17 06:20:38 +00:00
Reid Spencer
01cc8f3c76
Remove insignificant test no longer needed.
...
llvm-svn: 39931
2007-07-16 22:03:36 +00:00
Dan Gohman
0c1259b7e4
Remove the trailing semicolon from function declarations in LLVM,
...
where it's interpreted as a comment, not part of the syntax.
llvm-svn: 39920
2007-07-16 13:37:30 +00:00
Chris Lattner
96a035fce7
rename function to avoid llvm-upgrade warning
...
llvm-svn: 39895
2007-07-16 04:09:00 +00:00