Commit Graph

192 Commits

Author SHA1 Message Date
Dan Gohman ef3ef7f645 Fix GVN's debug output, now that operator<< on Value* doesn't print
a trailing newline.

llvm-svn: 77719
2009-07-31 20:24:18 +00:00
Owen Anderson 23a204d91b Move getTrue() and getFalse() to 2.5-like APIs.
llvm-svn: 77685
2009-07-31 17:39:07 +00:00
Owen Anderson b292b8ce70 Move more code back to 2.5 APIs.
llvm-svn: 77635
2009-07-30 23:03:37 +00:00
Dan Gohman 1ddf98ad8e Convert a few more things to use raw_ostream.
llvm-svn: 77039
2009-07-25 01:43:01 +00:00
Daniel Dunbar 0dd5e1ed39 More migration to raw_ostream, the water has dried up around the iostream hole.
- Some clients which used DOUT have moved to DEBUG. We are deprecating the
   "magic" DOUT behavior which avoided calling printing functions when the
   statement was disabled. In addition to being unnecessary magic, it had the
   downside of leaving code in -Asserts builds, and of hiding potentially
   unnecessary computations.

llvm-svn: 77019
2009-07-25 00:23:56 +00:00
Owen Anderson 47db941fd3 Get rid of the Pass+Context magic.
llvm-svn: 76702
2009-07-22 00:24:57 +00:00
Owen Anderson c37bc69e91 Rename getConstantInt{True|False} to get{True|False} at Chris' behest.
llvm-svn: 76598
2009-07-21 18:03:38 +00:00
Torok Edwin fbcc663cbf llvm_unreachable->llvm_unreachable(0), LLVM_UNREACHABLE->llvm_unreachable.
This adds location info for all llvm_unreachable calls (which is a macro now) in
!NDEBUG builds.
In NDEBUG builds location info and the message is off (it only prints
"UREACHABLE executed").

llvm-svn: 75640
2009-07-14 16:55:14 +00:00
Torok Edwin 56d0659726 assert(0) -> LLVM_UNREACHABLE.
Make llvm_unreachable take an optional string, thus moving the cerr<< out of
line.
LLVM_UNREACHABLE is now a simple wrapper that makes the message go away for
NDEBUG builds.

llvm-svn: 75379
2009-07-11 20:10:48 +00:00
Owen Anderson 1e5f00e7a7 This started as a small change, I swear. Unfortunately, lots of things call the [I|F]CmpInst constructors. Who knew!?
llvm-svn: 75200
2009-07-09 23:48:35 +00:00
Nick Lewycky a21d3daadc Remove the vicmp and vfcmp instructions. Because we never had a release with
these instructions, no autoupgrade or backwards compatibility support is
provided.

llvm-svn: 74991
2009-07-08 03:04:38 +00:00
Owen Anderson b5618da226 Convert the first batch of passes to use LLVMContext.
llvm-svn: 74748
2009-07-03 00:17:18 +00:00
Dale Johannesen 81b6463ed0 This fixes a bug introduced in 72661, which can
move loads back past a check that the load address
is valid, see new testcase.  The test that went
in with 72661 has exactly this case, except that
the conditional it's moving past is checking
something else; I've settled for changing that
test to reference a global, not a pointer.  It
may be possible to scan all the tests you pass and
make sure none of them are checking any component
of the address, but it's not trivial and I'm not
trying to do that here.

llvm-svn: 73632
2009-06-17 20:48:23 +00:00
Torok Edwin ba93ea7632 Add debug message about non-local loads being clobbered.
llvm-svn: 73625
2009-06-17 18:48:18 +00:00
Dan Gohman a8f8a85388 Make the EnableLoadPRE variable static.
llvm-svn: 73398
2009-06-15 18:30:15 +00:00
Dan Gohman a5b9645c4b Split the Add, Sub, and Mul instruction opcodes into separate
integer and floating-point opcodes, introducing
FAdd, FSub, and FMul.

For now, the AsmParser, BitcodeReader, and IRBuilder all preserve
backwards compatability, and the Core LLVM APIs preserve backwards
compatibility for IR producers. Most front-ends won't need to change
immediately.

This implements the first step of the plan outlined here:
http://nondot.org/sabre/LLVMNotes/IntegerOverflow.txt

llvm-svn: 72897
2009-06-04 22:49:04 +00:00
Owen Anderson cc0c75c74d Be more aggressive in doing LoadPRE by tracing backwards when a block only has
a single predecessor.

Patch by Jakub Staszak.

llvm-svn: 72661
2009-05-31 09:03:40 +00:00
Bill Wendling 006459ecd4 Enable GVN Load PRE.
llvm-svn: 72589
2009-05-29 20:38:16 +00:00
Torok Edwin 0b0ddb21fe just show the instruction, its not that slow.
llvm-svn: 72577
2009-05-29 16:58:36 +00:00
Torok Edwin 6a94624a1b for instructions with void type we have no choice but print the instruction as
is, otherwise we get a <badref>.

llvm-svn: 72567
2009-05-29 10:28:44 +00:00
Torok Edwin 72070282eb Add a DEBUG() output to GVN that prints the instruction clobbering a load.
This is useful when trying to figure out why GVN didn't eliminate redundant
loads.

llvm-svn: 72565
2009-05-29 09:46:03 +00:00
Owen Anderson 04cfdd38a2 Fix an issue where phiMap was not being updated properly when doing load PRE.
Diagnosis and patch thanks to Jakub Staszak.

llvm-svn: 72562
2009-05-29 05:37:54 +00:00
Duncan Sands 1efabaaa2a Allow readonly functions to unwind exceptions. Teach
the optimizers about this.  For example, a readonly
function with no uses cannot be removed unless it is
also marked nounwind.

llvm-svn: 71071
2009-05-06 06:49:50 +00:00
Owen Anderson 98f912bf13 Reapply r68211, with the miscompilations it caused fixed.
llvm-svn: 68262
2009-04-01 23:53:49 +00:00
Dan Gohman c4971721ea Revert r68172. It caused regressions in
Applications/Burg/burg
  Applications/ClamAV/clamscan
and many other tests.

llvm-svn: 68211
2009-04-01 16:37:47 +00:00
Owen Anderson ff5961b46c Enhance GVN to propagate simple conditionals. This fixes PR3921.
llvm-svn: 68172
2009-04-01 01:20:45 +00:00
John Criswell 073e4d16c5 Do not attempt to do parial redundancy elimination on void values.
Also fixed a punctuation error in the header comment.
This fixes PR3775.

llvm-svn: 66542
2009-03-10 15:04:53 +00:00
Devang Patel e8c6d3102d Skip DbgInfoIntrinsic.
llvm-svn: 66244
2009-03-06 02:59:27 +00:00
Chris Lattner 096f44de61 improve naming of values in GVN, patch by Jay Foad!
llvm-svn: 64363
2009-02-12 07:00:35 +00:00
Bill Wendling 415515077b Mistakenly turned this on.
llvm-svn: 64065
2009-02-08 01:32:00 +00:00
Bill Wendling 5469ec1072 Revert r63999. It was breaking self-hosting builds.
llvm-svn: 64062
2009-02-08 00:58:05 +00:00
Chris Lattner 73d7fe5a34 improve compatibility with cygwin, patch by Jay Foad!
llvm-svn: 62535
2009-01-19 22:00:18 +00:00
Owen Anderson 164274eeb1 Don't forget to remove phi nodes from the value numbering table after we collapse them.
llvm-svn: 61358
2008-12-23 00:49:51 +00:00
Bill Wendling 456e885382 Comment clean-ups. No functionality change.
llvm-svn: 61354
2008-12-22 22:32:22 +00:00
Bill Wendling e7f08e7250 Check that the instruction isn't in the value numbering scope.
llvm-svn: 61353
2008-12-22 22:28:56 +00:00
Bill Wendling 86f01cb9f6 Simplification: Negate the operator== method instead of implementing a full operator!= method.
llvm-svn: 61352
2008-12-22 22:16:31 +00:00
Bill Wendling 3c793441cb Add verification that deleted instruction isn't hiding in the PHI map.
llvm-svn: 61350
2008-12-22 22:14:07 +00:00
Bill Wendling ebb6a543fa Verify removed in a few more places.
llvm-svn: 61349
2008-12-22 21:57:30 +00:00
Bill Wendling 6b18a3994b Add verification functions to GVN which check to see that an instruction was
truely deleted. These will be expanded with further checks of all of the data
structures.

llvm-svn: 61347
2008-12-22 21:36:08 +00:00
Bill Wendling 070de29fcf Didn't mean to commit this.
llvm-svn: 61222
2008-12-18 22:19:50 +00:00
Bill Wendling 4c13e77d49 Re-XFAIL this test until debug stuff settles down.
llvm-svn: 61219
2008-12-18 22:13:31 +00:00
Chris Lattner b6372933b5 reapply this hunk from Bill's reversion in r61169, it is conservative
and safe and orthogonal from turning off load pre.

llvm-svn: 61177
2008-12-18 00:51:32 +00:00
Bill Wendling be4fb8a25f Temporarily revert r61027. It was causing a bootstrap failure in "release" mode
with everyone's favorite error messages:

Comparing stages 2 and 3
warning: ./cc1-checksum.o differs
warning: ./cc1plus-checksum.o differs
Bootstrap comparison failure!
./c-decl.o differs
./cp/decl.o differs
./df-core.o differs
./gcc.o differs
./i386.o differs
./stor-layout.o differs
./tree-pretty-print.o differs
./tree.o differs
make[2]: *** [compare] Error 1
make[1]: *** [stage3-bubble] Error 2

See PR3227.

llvm-svn: 61169
2008-12-17 23:31:20 +00:00
Chris Lattner 0c68ae0603 Enable Load PRE. This teaches GVN to push partially redundant loads up the
CFG when there is exactly one predecessor where the load is not available.
This is designed to not increase code size but still eliminate partially
redundant loads.  This fires 1765 times on 403.gcc even though it doesn't
do critical edge splitting yet (the most common reason for it to fail).

llvm-svn: 61027
2008-12-15 05:28:29 +00:00
Owen Anderson 03aacbae90 Ifdef out some code that I didn't mean to enable by default yet.
llvm-svn: 61024
2008-12-15 03:52:17 +00:00
Chris Lattner 69131fd872 make GVN try to rename inputs to the resultant replaced values, which
cleans up the generated code a bit.  This should have the added benefit of
not randomly renaming functions/globals like my previous patch did. :)

llvm-svn: 61023
2008-12-15 03:46:38 +00:00
Owen Anderson bfe133e4ac Add support for slow-path GVN with full phi construction for scalars. This is disabled for now, as it actually pessimizes code in the abscence
of phi translation for load elimination.  This slow down GVN a bit, by about 2% on 403.gcc.

llvm-svn: 61021
2008-12-15 02:03:00 +00:00
Owen Anderson e34c2399de Generalize GVN's phi construciton routine to work for things other than loads.
llvm-svn: 61009
2008-12-14 19:10:35 +00:00
Bill Wendling 293b9181e5 Temporarily revert r60973. It's inexplicably causing a failure when self-hosting LLVM:
llvm[2]: Linking Release executable opt (without symbols)
...
Undefined symbols:
  "llvm::APFloat::IEEEsingle", referenced from:
      __ZN4llvm7APFloat10IEEEsingleE$non_lazy_ptr in libLLVMCore.a(Constants.o)
      __ZN4llvm7APFloat10IEEEsingleE$non_lazy_ptr in libLLVMCore.a(AsmWriter.o)
      __ZN4llvm7APFloat10IEEEsingleE$non_lazy_ptr in libLLVMCore.a(ConstantFold.o)
  "llvm::APFloat::IEEEdouble", referenced from:
      __ZN4llvm7APFloat10IEEEdoubleE$non_lazy_ptr in libLLVMCore.a(Constants.o)
      __ZN4llvm7APFloat10IEEEdoubleE$non_lazy_ptr in libLLVMCore.a(AsmWriter.o)
      __ZN4llvm7APFloat10IEEEdoubleE$non_lazy_ptr in libLLVMCore.a(ConstantFold.o)
ld: symbol(s) not found

This is in release mode. To replicate, compile llvm and llvm-gcc in optimized
mode. Then build llvm, in optimized mode, with the newly created compiler.

llvm-svn: 60977
2008-12-13 09:28:44 +00:00
Chris Lattner 1e29f7c97d make RLE preserve the name of the load that it replaces. This is just
a pretification of the IR.

llvm-svn: 60973
2008-12-13 07:22:47 +00:00