Commit Graph

2056 Commits

Author SHA1 Message Date
Owen Anderson
891eecb040 Avoid potential iterator invalidation problems.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40607 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-30 21:26:39 +00:00
Devang Patel
df5cf2074c Remove dead code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40606 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-30 21:10:44 +00:00
Devang Patel
d9a6dcba9f Loop Rotation pass preserves dominator tree and frontier.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40603 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-30 20:22:53 +00:00
Devang Patel
96b651c627 LICM preserves scalar evolution and dom frontier.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40602 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-30 20:19:59 +00:00
Reid Spencer
ed3fa8552a Fix a typo/thinko.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40599 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-30 19:53:57 +00:00
Owen Anderson
45c8388e2a Use more caching when computing non-local dependence. This makes bzip2 not
use up the entire 32-bit address space.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40596 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-30 17:29:24 +00:00
Owen Anderson
4f9ba7c40c Fix a bug caused by indiscriminantly asking for the dominators of a predecessor.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40595 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-30 16:57:08 +00:00
Chuck Rose III
936baaa5ae VStudio compiler errors and placing Function*->ExFunc map under ManagedStatic control.
This commit fixes two things.  One is a pair of VStudio compiler errors stemming from variables
which defined within the for loop statement and also within the body of the for loop.  I fixed these 
by renaming one of the two variables.  Additionally, I've made the Function*->ExFunc map in 
ExternalFunctions.cpp a ManagedStatic object, so that cleanup will be done on llvm_shutdown.  In repeated
uses of the interpreter, where the same Function* address may get used for completely differnet functions,
this was causing a crash.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40558 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-27 18:26:35 +00:00
Devang Patel
70b36d96b3 Fix thinko. Update return status appropriately.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40546 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-26 20:21:42 +00:00
Owen Anderson
45537917ee Fix a couple more bugs in the phi construction by pulling in code that does
almost the same things from LCSSA.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40540 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-26 18:26:51 +00:00
Dan Gohman
c2dfd066c0 Remove a bogus return statement, what appears to have been a pasto
from Relation::contradicts in Relation::incorporate.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40531 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-26 15:29:35 +00:00
Owen Anderson
febc7e3613 Fix what is _hopefully_ the last corner case for loops.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40503 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-25 23:54:42 +00:00
Owen Anderson
9f435740b1 My last commit was not correct for nested loops. Fix it, and add a testcase for it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40498 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-25 22:19:40 +00:00
Owen Anderson
0d16988690 Fix an infinite loop on 300.twolf.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40497 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-25 22:03:06 +00:00
Owen Anderson
2e3d6b4a27 Fix a bug that was causing GVN to crash on 252.eon.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40494 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-25 21:13:41 +00:00
Owen Anderson
0cd320362e Add basic support for performing whole-function RLE.
Note: This has not yet been thoroughly tested.  Use at your own risk.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40489 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-25 19:57:03 +00:00
Owen Anderson
1ad2cb7555 Add a GVN pass, using the value numbering code I developed for GVNPRE and the
load elimination code from RedundantLoadElimination.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40469 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-24 17:55:58 +00:00
Owen Anderson
a7f98e2919 Rename a lot of things to change FastDLE to RedundantLoadElimination.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40457 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-24 00:17:04 +00:00
Owen Anderson
e2a1624da3 Rename FastDLE as RedundantLoadElimination.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40456 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-24 00:08:38 +00:00
Owen Anderson
6b8894bb67 Don't delete volatile loads. Doing so is not safe.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40448 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-23 22:05:54 +00:00
Owen Anderson
ffe40aa97d Add FastDLE, the load-elimination counterpart of FastDSE.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40445 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-23 21:48:08 +00:00
Owen Anderson
d989711844 Fix file header.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40440 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-23 18:30:37 +00:00
Chris Lattner
410ce1c49a completely remove a transformation that is unsafe in the face of
undefs.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40439 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-23 17:10:17 +00:00
Devang Patel
00ae1a2f02 Apply temporary work around to fix llvm mis-compilation
reported in PR 1556.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40133 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-21 00:34:29 +00:00
Chris Lattner
3ccc6bc61c this xform is already done by the constant folder.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40124 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-20 22:06:41 +00:00
Dan Gohman
9941f7426d Optimize alignment of loads and stores.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40102 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-20 16:34:21 +00:00
Duncan Sands
e2abf12436 Place SCCPSolver also in the anonymous namespace. This
pacifies g++-4.2.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40089 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-20 08:56:21 +00:00
Owen Anderson
c9399be5e5 Fix a bug where we were marking GEP expressions with the wrong opcode.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40085 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-20 08:19:20 +00:00
Owen Anderson
b83e56f6b6 Make val_replace fail early, which reduces the time to optimize 403.gcc to 14.8s.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40064 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-19 19:57:13 +00:00
Owen Anderson
19bc4a8acd Use SmallVector and DenseMap in even more places.
With this, the time to optimize 403.gcc is down to 15.1s.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40042 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-19 06:37:56 +00:00
Owen Anderson
fb66541e63 Change ValueTable to use a DenseMap for mapping expressions to value numbers.
This results in a slight speedup for 403.gcc.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40040 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-19 06:13:15 +00:00
Owen Anderson
b97819852a Move some sets and maps to SmallPtrSet and DenseMap respectively. This
reduces the time to optimize 403.gcc from 17.6s to 16.4s.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40036 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-19 03:32:44 +00:00
Devang Patel
24857a3fac Fix typo.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40025 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-18 23:50:19 +00:00
Devang Patel
31ad75329e Fix dominator info update to accommodate CFG changes.
This fixes PR1559.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40024 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-18 23:48:20 +00:00
Owen Anderson
ce17b1c1f8 Take advantage of undefined behavior if the source program tries to GEP
beyond the end of an alloca to make FastDSE faster and more aggressive.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@39945 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-16 23:34:39 +00:00
Owen Anderson
faac518ce0 Add support for walking up memory def chains, which enables finding many more
dead stores on 400.perlbench.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@39929 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-16 21:52:50 +00:00
Dan Gohman
07a96765da Fix comments about vectors to use the current wording.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@39921 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-16 14:29:03 +00:00
Chris Lattner
cb7122bb7d Repair a regression in Transforms/InstCombine/mul.ll that Reid noticed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@39896 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-16 04:15:34 +00:00
Nick Lewycky
5380e941ef Start adding and cleaning up comments.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@39894 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-16 02:58:37 +00:00
Chris Lattner
4241e4da85 Implement shift-simplify.ll:test[45].
First teach instcombine that sign bit checks only demand the 
sign bit, this allows simplify demanded bits to hack on 
expressions better.

Second, teach instcombine that ashr is useless if only the
sign bit is demanded.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@39880 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-15 20:54:51 +00:00
Chris Lattner
a0141b90a3 Implement shift-simplify.ll:test3, turning:
(X << 31) <s 0  --> (X&1) != 0

This happens dozens of times in the CFE.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@39879 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-15 20:42:37 +00:00
Nick Lewycky
a73d11efd4 Use maximal intersection algorithm exclusively. Fixes miscompile bug.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@39852 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-14 04:28:04 +00:00
Owen Anderson
a08971559d Handle GEPs with all-zero indices in the same way we handle pointer-pointer bitcasts. Also, fix a potentia infinite loop.
This brings FastDSE to parity with old DSE on 175.vpr.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@39839 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-13 22:50:48 +00:00
Devang Patel
fba36b8cba Disable claims to preserve analysis until open issues are resolved.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@39834 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-13 21:53:42 +00:00
Owen Anderson
3e8d001799 Be more aggressive in removing dead stores, and in removing instructions trivially dead after DSE.
This drastically improves the effect of FastDSE on kimwitu++.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@39819 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-13 18:26:26 +00:00
Owen Anderson
43b2676cc2 Reimplement removing stores to allocas at the end of a function. This should be safe now.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@39790 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-12 21:41:30 +00:00
Owen Anderson
dd61c2b25c Make the condition-checking for free with non-trivial dependencies more correct.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@39789 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-12 18:08:51 +00:00
Owen Anderson
1b4542b6df Remove the end-block handling code. It was unsafe, and making it safe would have resulted in falling back to the slow DSE case. I need to think some more about the right way to handle this.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@39788 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-12 17:52:20 +00:00
Gabor Greif
3b0a7ea66e checked in as obvious,
thanks Benoit Boissinot!


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@39774 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-12 13:31:38 +00:00
Owen Anderson
1cb960a4bb Let MemoryDependenceAnalysis take care of updating AliasAnalysis.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@39769 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-12 00:06:21 +00:00
Devang Patel
990e866deb Preserve analysis info.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@39767 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-11 23:47:28 +00:00
Owen Anderson
a96c1f665a Handle the case where an entire structure is freed, and its dependency is a store to a field within
that structure.

Also, refactor the runOnBasicBlock() function, splitting some of the special cases into separate functions.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@39762 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-11 23:19:17 +00:00
Owen Anderson
6f46d65104 Add support for eliminate stores to stack-allocated memory locations at the end
of a function.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@39754 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-11 21:06:56 +00:00
Owen Anderson
2655adb3b2 Handle eliminating stores that occur right before a free.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@39753 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-11 20:38:34 +00:00
Owen Anderson
bfbfb3c464 Clean up a few things based on Chris' feedback.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@39747 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-11 19:03:09 +00:00
Owen Anderson
b77c457cc8 Add FastDSE, a new algorithm for doing dead store elimination. This algorithm is not as accurate
as the current DSE, but it only a linear scan over each block, rather than quadratic.  Eventually
(once it has been improved somewhat), this will replace the current DSE.

NOTE: This has not yet been extensively tested.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@38517 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-11 00:46:18 +00:00
Owen Anderson
16a72bb47d Make the pass registration static.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@38508 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-10 20:20:19 +00:00
Nick Lewycky
6e8eb7f075 Update the ValueRanges interface to use value numbers instead of Value*s.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@38483 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-10 03:28:21 +00:00
Owen Anderson
81c2a6ecbb Move some key maps from std::map to DenseMap. This improves the time to optimize Anton's testcase from 17.5s
to 15.7s.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@38480 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-10 00:27:22 +00:00
Owen Anderson
a05a81b10a Use a cheaper test, delaying calling find_leader() until we know that it's necessary. This improves
the time to optimize Anton's testcase from 21.1s to 17.6s.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@38479 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-10 00:09:25 +00:00
Owen Anderson
5fd507d431 Add an assertion if find_leader fails.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@38477 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-09 23:57:18 +00:00
Owen Anderson
dfa2435784 Take advantage of the new fast SmallPtrSet assignment operator when propagating AVAIL_OUT sets.
This reduces the time to optimize Anton's testcase from 31.2s to 21.s!


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@38475 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-09 22:29:50 +00:00
Devang Patel
ff366850aa Expose struct size threhold to allow users to tweak their own setting.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@38472 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-09 21:19:23 +00:00
Owen Anderson
ab3fd054ea Fix a comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@38459 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-09 16:43:55 +00:00
Owen Anderson
b9eeb1acbc Improve a hotspot that was making build_sets() slower by calling lookup() too
often.  This improves Anton's testcase from 36s to 32s.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@38441 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-09 07:56:55 +00:00
Owen Anderson
0616dffd64 Start using a set representation that remembers the set of value numbers represented
in the set.  For the moment, this results in a slight performance decrease, but
it lays the groundwork for future improvements.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@38439 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-09 06:50:06 +00:00
Owen Anderson
36a07daabd Fix an error where ANTIC_OUT was ending up with more than one expression of
the same value number.  This fixes an infinite loop on 444.namd.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37967 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-07 20:13:57 +00:00
Nick Lewycky
8f5149e002 Back out Devang's fix for PR1320 because it causes PR1542.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37966 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-07 16:23:34 +00:00
Devang Patel
05c1dc6493 These rountines are now available as part of basic block utilities.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37955 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-06 22:03:47 +00:00
Devang Patel
7be06f9158 Request DominanceFrontiner in advance.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37954 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-06 21:43:22 +00:00
Devang Patel
658f625d16 Preserve various analysis info.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37953 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-06 21:40:13 +00:00
Owen Anderson
ec5614b522 Be more aggressive in the heuristic. This mostly exposes more opportunities
for the GVN part of GVNPRE to apply.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37951 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-06 20:29:43 +00:00
Owen Anderson
2c194e6bd1 Achieve what the incorrect test was trying to do by simply requiring that all
critical edges be split before we begin.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37949 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-06 18:12:36 +00:00
Owen Anderson
6a6152ca2d Remove an incorrect check.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37948 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-06 16:52:47 +00:00
Zhou Sheng
33b0b8d242 Correct a typo.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37936 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-06 06:01:16 +00:00
Owen Anderson
2a5c9d8aac Fix a bunch of issues found in a testcase from 400.perlbench.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37929 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-05 23:11:26 +00:00
Nick Lewycky
29a05b6a71 Break "variable canonicalization" out of InequalityGraph and into its own class
"ValueNumbering".


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37881 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-05 03:15:00 +00:00
Owen Anderson
ec3ed5f8e4 Fix another bug, this time in PREing select instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37878 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-04 22:33:23 +00:00
Owen Anderson
df30b63663 Fix a typo that was killing GVNPRE of select instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37871 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-04 18:26:18 +00:00
Owen Anderson
9eee94c5f7 Fix an error in phi translation of GEPs that was causing failures.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37868 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-04 04:51:16 +00:00
Owen Anderson
5653322e72 Add support for performing GVNPRE on GEP instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37862 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-03 23:51:19 +00:00
Owen Anderson
eb21686ea9 Add functionality to value number GEP instructions. This also provides the infrastructure that will
be used for function calls.  NOTE: This does not yet do any transformation of GEPs or function calls.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37860 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-03 22:50:56 +00:00
Owen Anderson
3d6fac3993 Make the unary operator case a bit faster, since casts are the only kind of unary operation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37857 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-03 19:01:42 +00:00
Owen Anderson
216394f9b3 Add support for performing GVNPRE on cast instructions, and add a testcase for this.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37856 91177308-0d34-0410-b5e6-96231b3b80d8
2007-07-03 18:37:08 +00:00
Devang Patel
f34a43acc0 Preserve DominanceFrontier.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37820 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-29 23:11:49 +00:00
Devang Patel
6c631988c7 Do not filter loop if candidate branch is in loop header.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37792 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-29 01:39:53 +00:00
Owen Anderson
ca6c31cc41 Add support for value numbering (but not actually optimizing) cast instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37789 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-29 00:51:03 +00:00
Owen Anderson
40dc00eb82 Add a type field to expressions in preparation for performing GVNPRE on casts.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37788 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-29 00:40:05 +00:00
Owen Anderson
890e1a0401 Add support for performing GVNPRE on select instructions. This fixes test/Transforms/GVNPRE/select.ll.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37783 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-28 23:51:21 +00:00
Devang Patel
eb62eca503 - Undo previous check and allow loop switch for condtion that is not inside
loop.
- Avoid loop unswich for loop header branch.
- While cloning dominators fix typo and handle self dominating blocks.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37772 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-28 02:05:46 +00:00
Devang Patel
cce624a7a9 Update LoopUnswitch pass to preserve DomiantorTree.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37771 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-28 00:49:00 +00:00
Devang Patel
558f1b8439 If a condition is not inside a loop then the condition is suitable
to loop unswitch candidate for the loop.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37770 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-28 00:44:10 +00:00
Owen Anderson
a20f35d2e7 Make many sets a much more reasonable size. This decreases the time to optimize
Anton's testcase from 35.5s to 34.7s.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37769 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-28 00:34:34 +00:00
Owen Anderson
68cb52e468 Use cached information that has already been computed to make clean() simpler and faster. This is a small speedup on most cases.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37761 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-27 17:38:29 +00:00
Owen Anderson
7b317d2f59 Fold a lot of code into two cases: binary instructions and ternary instructions.
This saves many lines of code duplication.  No functionality change.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37759 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-27 17:03:03 +00:00
Zhou Sheng
0d7d3656f3 Fix a bug.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37751 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-27 09:50:26 +00:00
Owen Anderson
62cf8babdb Add support for performing GVNPRE on the three vector-specific operations.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37745 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-27 04:10:46 +00:00
Owen Anderson
6032a5ba64 1. Correct some comments and clean up some dead code.
2. When calculating ANTIC_IN, only iterate the changed blocks.  For most average
inputs this is a small speedup, but for cases with unusual CFGs, this can be a significant win.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37742 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-26 23:29:41 +00:00
Owen Anderson
9030d384c4 Use the built-in postorder iterators rather than computing a postorder walk by hand.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37721 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-25 18:25:31 +00:00
Owen Anderson
f62c44a38d 1) Fix an issue with non-deterministic iteration order in phi_translate
2) Remove some maximal-set computing code that is no longer used.
3) Use a post-order CFG traversal to compute ANTIC_IN instead of a postdom traversal.
This causes the ANTIC_IN calculation to converge much faster.  Thanks to Daniel Berlin for suggesting this.

With this patch, the time to optimize 403.gcc decreased from 17.5s to 7.5s, and Anton's huge
testcase decreased from 62 minutes to 38 seconds.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37714 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-25 05:41:12 +00:00
Nick Lewycky
4390feb7ae Fix value ranges.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37713 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-24 20:14:22 +00:00
Owen Anderson
79a6370133 Fix a silly mistake that was causing failures.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37712 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-24 08:42:24 +00:00
Nick Lewycky
dea252689d Remove tabs.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37710 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-24 04:40:16 +00:00
Nick Lewycky
984504b912 Remove use of ETForest. Also cleaned up issues around unreachable basic
blocks, and optimizing within one basic block.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37709 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-24 04:36:20 +00:00
Owen Anderson
647580483c Rework topo_sort so eliminate some behavior that scaled terribly. This reduces the time to optimize 403.gcc from 18.2s to 17.5s,
and has an even larger effect on larger testcases.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37708 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-22 21:31:16 +00:00
Owen Anderson
2106f61f23 Perform fewer set insertions while calculating ANTIC_IN. This reduces the amount of time to optimize 403.gcc from 21.9s to 18.2s.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37707 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-22 18:27:04 +00:00
Owen Anderson
b3b37345de Remove some code that I was using for collecting performance information that should not have been committed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37706 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-22 17:04:40 +00:00
Owen Anderson
124084604d Avoid excessive calls to find_leader when calculating AVAIL_OUT. This reduces the time to optimize 403.gcc from 23.5s to 21.9s.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37702 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-22 03:14:03 +00:00
Owen Anderson
e8138ffeaf Reserve space in vectors before topologically sorting into them. This improves the time to optimize 403.gcc from 28s to 23.5s.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37699 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-22 00:43:22 +00:00
Owen Anderson
82575d8ab1 Make a bunch of optimizations for compile time to GVNPRE, including smarter set unions, deferring blocks rather than computing maximal sets, and smarter use of sets. With these enhancements, the time to optimize 273.perlbmk goes from 5.3s to 2.7s.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37698 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-22 00:20:30 +00:00
Chris Lattner
1dbfd48fa2 Significantly improve the documentation of the instcombine divide/compare
transformation.  Also, keep track of which end of the integer interval overflows
occur on.  This fixes Transforms/InstCombine/2007-06-21-DivCompareMiscomp.ll
and rdar://5278853, a miscompilation of perl.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37692 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-21 18:11:19 +00:00
Owen Anderson
66fd906074 Change lots of sets from std::set to SmallPtrSet. This reduces the time required to optimize 253.perlbmk from 10.9s to 5.3s.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37690 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-21 17:57:53 +00:00
Owen Anderson
58e087b6e3 Eliminate a redundant check. This speeds up optimization of 253.perlbmk from 13.5 seconds to 10.9 seconds.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37683 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-21 01:59:05 +00:00
Owen Anderson
9cb56014ce Comment-ize the functions in GVNPRE.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37681 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-21 00:19:05 +00:00
Chris Lattner
562ef78df2 refactor a bunch of code out of visitICmpInstWithInstAndIntCst into its own
routine.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37679 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-20 23:46:26 +00:00
Owen Anderson
3eaca716bf Split runOnFunction into many smaller functions. This make it easier to get accurate performance analysis of GVNPRE.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37678 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-20 22:10:02 +00:00
Owen Anderson
0304b2bc2b Make GVNPRE accurate report whether it modified the function or not.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37673 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-20 18:30:20 +00:00
Owen Anderson
5dbea73e35 Get rid of an unneeded helper function.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37670 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-20 00:43:33 +00:00
Owen Anderson
4f703ec0a1 Use a DenseMap instead of an std::map for the value numbering. This reduces the time to optimize lencod on a PPC Debug build from ~300s to ~140s.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37668 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-19 23:23:54 +00:00
Owen Anderson
52471b102c Make dependsOnInvoke much more specific in what it tests, which in turn make it much faster to run. This reduces the time to optimize lencondwith a debug build on PPC from ~450s to ~300s.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37667 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-19 23:07:16 +00:00
Dan Gohman
5cec4db6ae Rename ScalarEvolution::deleteInstructionFromRecords to
deleteValueFromRecords and loosen the types to all it to accept
Value* instead of just Instruction*, since this is what
ScalarEvolution uses internally anyway. This allows more flexibility
for future uses.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37657 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-19 14:28:31 +00:00
Owen Anderson
d0954105f0 Handle constants in phi nodes properly. This fixes test/Transforms/GVNPRE/2007-06-18-ConstantInPhi.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37655 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-19 07:35:36 +00:00
Chris Lattner
2384d7b179 silence a bogus warning Duraid ran into.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37649 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-19 05:43:49 +00:00
Owen Anderson
20cb51fda4 Be careful to erase values from all of the appropriate sets when they're not needed anymore. This fixes a few more memory-related issues.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37647 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-19 05:37:32 +00:00
Owen Anderson
6cae7409b1 Remember to clear the maximal sets between functions.
Thanks to Nicholas for valgrinding this.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37646 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-19 04:32:55 +00:00
Owen Anderson
086f5f350b Refactor GVNPRE to use a much smart method of uniquing value sets, and centralize a lot of the value numbering information. No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37645 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-19 03:31:41 +00:00
Owen Anderson
bbf1197be1 Cache the results of dependsOnInvoke()
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37622 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-18 04:42:29 +00:00
Owen Anderson
a8daa7176b Fix indentation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37621 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-18 04:31:21 +00:00
Owen Anderson
60e174495a Don't perform an expensive check if it's not necessary.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37620 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-18 04:30:44 +00:00
Owen Anderson
32bc789b6c Fix test/Transforms/GVNPRE/2007-06-15-InvokeInst.ll by ignoring all instructions that depend on invokes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37610 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-16 00:26:54 +00:00
Dan Gohman
e19dd87598 Use SCEVConstant::get instead of SCEVUnknown::get to create an
integer constant SCEV.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37596 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-15 18:00:55 +00:00
Owen Anderson
383bcb2e78 Fix test/Transforms/GVNPRE/2007-06-15-Looping.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37595 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-15 17:55:15 +00:00
Dan Gohman
d19534add9 Add a SCEV class and supporting code for sign-extend expressions.
This created an ambiguity for expandInTy to decide when to use
sign-extension or zero-extension, but it turns out that most of its callers
don't actually need a type conversion, now that LLVM types don't have
explicit signedness. Drop expandInTy in favor of plain expand, and change
the few places that actually need a type conversion to do it themselves.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37591 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-15 14:38:12 +00:00
Chris Lattner
7cbe2ebb21 Generalize many transforms to work on ~ of vectors in addition to ~ of
integer ops.  This implements Transforms/InstCombine/and-or-not.ll
test3/test4, and finishes off PR1510


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37589 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-15 06:23:19 +00:00
Chris Lattner
003b6200df Implement two xforms:
1. ~(~X | Y) === (X & ~Y)
2. (A|B) & ~(A&B) -> A^B

This allows us to transform  ~(~(a|b) | (a&b)) -> a^b.

This implements PR1510 for scalar values.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37584 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-15 05:58:24 +00:00
Chris Lattner
041a6c9048 delete some obviously dead vector operations, which deletes a few thousand
operations from Duraids example.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37582 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-15 05:26:55 +00:00
Owen Anderson
71fcebcf8a Fix test/Transforms/GVNPRE/2007-06-12-PhiTranslate.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37564 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-12 22:43:57 +00:00
Owen Anderson
239e71201e Refactor some code, and fix test/Transforms/GVNPRE/2007-06-12-NoExit.ll by being more careful when using
post-dominator information.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37556 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-12 16:57:50 +00:00
Dale Johannesen
ce0b23721e Sink CmpInst's to their uses to reduce register pressure.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37554 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-12 16:50:17 +00:00
Owen Anderson
65d2862a60 Fix a few more bugs, including an instance of walking in reverse topological rather than topological order. This
fixes a testcase extracted from llvm-test.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37550 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-12 00:50:47 +00:00
Owen Anderson
0e71466405 Handle functions with multiple exit blocks properly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37539 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-11 16:25:17 +00:00
Owen Anderson
139fe848a2 Perform PRE of comparison operators.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37536 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-09 18:35:31 +00:00
Owen Anderson
b8b873cb7d Collect statistics from GVN-PRE.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37530 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-08 22:02:36 +00:00
Owen Anderson
a7f916cf89 Fix typo in a comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37526 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-08 20:57:08 +00:00
Owen Anderson
8338ff5ff1 Fix a bug that was causing the elimination phase not to replace values when it should be.
With this patch, GVN-PRE now correctly optimizes the example from the thesis.

Many thanks to Daniel Berlin for helping me find errors in this.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37525 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-08 20:44:02 +00:00
Owen Anderson
900997ba3b Small bugfix, and const-ify some methods (Thanks, Bill).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37513 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-08 01:52:45 +00:00
Owen Anderson
397d405731 Add partial redundancy elimination.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37510 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-08 01:03:01 +00:00
Devang Patel
468502e41a Do not require ETForest. Now it is unused by LICM.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37502 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-07 22:21:15 +00:00
Devang Patel
326821ef12 Use DominatorTree instead of ETForest.
This allows faster immediate domiantor walk.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37500 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-07 21:57:03 +00:00
Devang Patel
b7d9dfc7ba Use DominatorTree instead of ETForest.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37499 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-07 21:42:15 +00:00
Devang Patel
fa51f9ae12 Use DominatorTree instead of ETForest.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37498 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-07 21:35:27 +00:00
Devang Patel
a97e040e05 Use DominatorTree instead of ETForest.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37495 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-07 18:45:06 +00:00
Devang Patel
387b8cbc2d Use DominatorTree instead of ETForest.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37494 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-07 18:40:55 +00:00
Lauro Ramos Venancio
d0499af703 Fix PR1499.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37472 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-06 17:08:48 +00:00
Nick Lewycky
5adcefb04d Inform ScalarEvolutions that we're deleting Values.
This is the obviously correct part of the fix for PR1487.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37457 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-06 03:51:56 +00:00
Owen Anderson
12112af5e8 Add simple full redundancy elimination.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37455 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-06 01:27:49 +00:00
Chris Lattner
b47f6124f4 Fix PR1495 and CodeGen/X86/2007-06-05-LSR-Dominator.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37454 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-06 01:23:55 +00:00
Devang Patel
743f7e8839 Avoid non-trivial loop unswitching while optimizing for size.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37446 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-06 00:21:03 +00:00
Owen Anderson
1d4eb6ab0d Fix a misunderstanding of the algorithm. Really, we should be tracking values
and expression separately.  We can get around this, however, by only keeping
opaque values in TMP_GEN.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37443 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-05 23:46:12 +00:00
Owen Anderson
8f862c80bc Don't leak memory.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37442 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-05 22:11:49 +00:00
Owen Anderson
359208032b Fix a small bug, some 80 cols violations, and add some more debugging output.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37436 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-05 17:31:23 +00:00
Dan Gohman
d8af90c1da Allow insertelement, extractelement, and shufflevector to be hoisted/sunk
by LICM.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37435 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-05 16:05:55 +00:00
Bill Wendling
851879c4cf Patches by Chuck Rose to unbreak V Studio builds.
Thanks Chuck!


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37428 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-04 23:52:59 +00:00
Devang Patel
170f2b9bef s/ETNode::getChildren/ETNode::getETNodeChildren/g
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37426 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-04 23:45:02 +00:00
Owen Anderson
dbc705b73f Don't use std::set_difference when the two sets are sorted differently. Compute
the difference manually instead.

This allows GVNPRE to produce correct analysis for the example in the GVNPRE
paper.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37425 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-04 23:34:56 +00:00
Owen Anderson
082fe7185e Fix a bunch of small bugs, and improve the debugging output significantly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37424 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-04 23:28:33 +00:00
Owen Anderson
5ea069fdda Make phi_translate correct.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37418 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-04 18:05:26 +00:00
Devang Patel
26042420d6 s/llvm::DominatorTreeBase::DomTreeNode/llvm::DomTreeNode/g
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37407 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-04 00:32:22 +00:00
Owen Anderson
0fa6b37c6f Don't use the custom comparator where it's not necessary.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37406 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-03 22:02:14 +00:00
Devang Patel
bec7647f98 s/DominatorTreeBase::Node/DominatorTreeBase:DomTreeNode/g
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37403 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-03 06:26:14 +00:00
Owen Anderson
10ac137d0b Remove an unused method.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37402 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-03 05:58:25 +00:00
Owen Anderson
a724ac1a73 There's no need to have an Expression class... Value works just as well! This simplifies a lot of code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37401 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-03 05:55:58 +00:00
Devang Patel
e7ae1a9dee Insert new instructions in AliasSet.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37390 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-01 22:15:31 +00:00
Owen Anderson
aa7d3351f0 clean() needs to process things in topological order.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37389 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-01 22:00:37 +00:00
Owen Anderson
8e21fb7df8 Fix Expression comparison, which in turn fixes a value numbering error.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37386 91177308-0d34-0410-b5e6-96231b3b80d8
2007-06-01 17:34:47 +00:00
Owen Anderson
243f348c2a Add a topological sort function.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37376 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-31 22:44:11 +00:00
Owen Anderson
8214d50338 Attempt to fix up phi_translate.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37366 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-31 00:42:15 +00:00
Devang Patel
f38ac5dd39 Fix typo.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37360 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-30 15:29:37 +00:00
Chris Lattner
39a1c04323 Fix Transforms/ScalarRepl/2007-05-29-MemcpyPreserve.ll and the second
half of PR1421, by not decimating structs with holes that are the source and
destination of a memcpy.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37358 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-30 06:11:23 +00:00
Owen Anderson
2c0c51593d Fix a typo
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37350 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-29 23:34:14 +00:00
Owen Anderson
e3072b292f Re-fix a bug, where I was now being too aggressive.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37348 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-29 23:26:30 +00:00
Owen Anderson
cdf8efde6a Use proper debugging facilities so other people don't have to look at my commented-out
debugging lines.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37347 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-29 23:15:21 +00:00
Owen Anderson
9b6bd12ef0 Comment debug code out that I accidentally uncommented last time.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37346 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-29 22:43:03 +00:00
Owen Anderson
b502bdb604 Add a place where I missed using the maximal set. Note that using the maximal
set this way is _SLOW_.  Somewhere down the line, I'll look at speeding it up.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37345 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-29 22:35:41 +00:00
Owen Anderson
ea12a06a81 Very first part of a GVN-PRE implementation. It currently performs a bunch of analysis, and nothing more. It is also quite slow for the moment. However,
it should give a sense of what's going on.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37343 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-29 21:53:49 +00:00
Chris Lattner
7139406707 Fix PR1446 by not scalarrepl'ing giant structures.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37326 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-24 18:43:04 +00:00
Chris Lattner
2b7e0ade62 fix a miscompilation when passing a float through varargs
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37297 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-23 01:17:04 +00:00
Chris Lattner
a2b18de4ba Fix Transforms/InstCombine/2007-05-18-CastFoldBug.ll, a bug that devastates
objc code due to the way the FE lowers objc message sends.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37256 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-19 06:51:32 +00:00
Chris Lattner
fb3e1190fc Handle negative strides much more optimally. This compiles X86/lsr-negative-stride.ll
into:

_t:
        movl 8(%esp), %ecx
        movl 4(%esp), %eax
        cmpl %ecx, %eax
        je LBB1_3       #bb17
LBB1_1: #bb
        cmpl %ecx, %eax
        jg LBB1_4       #cond_true
LBB1_2: #cond_false
        subl %eax, %ecx
        cmpl %ecx, %eax
        jne LBB1_1      #bb
LBB1_3: #bb17
        ret
LBB1_4: #cond_true
        subl %ecx, %eax
        cmpl %ecx, %eax
        jne LBB1_1      #bb
        jmp LBB1_3      #bb17

instead of:

_t:
        subl $4, %esp
        movl %esi, (%esp)
        movl 12(%esp), %ecx
        movl 8(%esp), %eax
        cmpl %ecx, %eax
        je LBB1_4       #bb17
LBB1_1: #bb.outer
        movl %ecx, %edx
        negl %edx
LBB1_2: #bb
        cmpl %ecx, %eax
        jle LBB1_5      #cond_false
LBB1_3: #cond_true
        addl %edx, %eax
        cmpl %ecx, %eax
        jne LBB1_2      #bb
LBB1_4: #bb17
        movl (%esp), %esi
        addl $4, %esp
        ret
LBB1_5: #cond_false
        movl %ecx, %edx
        subl %eax, %edx
        movl %eax, %esi
        addl %esi, %esi
        cmpl %ecx, %esi
        je LBB1_4       #bb17
LBB1_6: #cond_false.bb.outer_crit_edge
        movl %edx, %ecx
        jmp LBB1_1      #bb.outer


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37252 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-19 01:22:21 +00:00
Devang Patel
9f26f73179 Fix PR1431
Test case at Transformations/SCCP/2007-05-16-InvokeCrash.ll


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37185 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-17 22:10:15 +00:00
Chris Lattner
6b6aef8efc Fix Transforms/InstCombine/2007-05-14-Crash.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37057 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-15 00:16:00 +00:00
Dan Gohman
d9efdb2012 Correct a few comments.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37034 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-14 14:31:17 +00:00
Chris Lattner
7d8ed8a941 significantly improve debug output of lsr
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36996 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-11 22:40:34 +00:00
Dan Gohman
23d9d27c26 Fix typos.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36994 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-11 21:10:54 +00:00
Dan Gohman
c767844fc9 This patch extends the LoopUnroll pass to be able to unroll loops
with unknown trip counts. This is left off by default, and a
command-line option enables it. It also begins to separate loop
unrolling into a utility routine; eventually it might be made usable
from other passes.

It currently works by inserting conditional branches between each
unrolled iteration, unless it proves that the trip count is a
multiple of a constant integer > 1, which it currently only does in
the rare case that the trip count expression is a Mul operator with
a ConstantInt operand. Eventually this information might be provided
by other sources, for example by a pass that peels/splits the loop
for this purpose.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36990 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-11 20:53:41 +00:00
Chris Lattner
3aea1bdca5 fix regressions from my previous checking, including
Transforms/InstCombine/2006-12-08-ICmp-Combining.ll


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36989 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-11 16:58:45 +00:00
Chris Lattner
8885887b9c fix Transforms/InstCombine/2007-05-10-icmp-or.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36984 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-11 05:55:56 +00:00
Devang Patel
fd98dc9a76 Fix PR1333
Testcases :
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070507/049451.html
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070507/049452.html


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36955 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-09 08:24:12 +00:00
Dan Gohman
1cc0096d55 Fix various whitespace inconsistencies.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36936 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-08 15:19:19 +00:00
Dan Gohman
f742c2363c Correct the comment for ApproximateLoopSize to reflect what it actually does.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36935 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-08 15:14:19 +00:00
Dale Johannesen
6603a1bff0 Don't generate branch to entry block.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36917 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-08 01:01:04 +00:00
Nick Lewycky
ecd94c804a Fix typo in comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36873 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-06 13:37:16 +00:00
Chris Lattner
6f6f5125e2 Fix a bug in my previous patch
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36857 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-06 07:24:03 +00:00
Chris Lattner
8c756c1fff Implement Transforms/InstCombine/cast_ptr.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36809 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-05 22:41:33 +00:00
Chris Lattner
d717c1859a wrap long lines
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36807 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-05 22:32:24 +00:00
Chris Lattner
caccc996e2 Fix Transforms/LoopUnroll/2007-05-05-UnrollMiscomp.ll and PR1385.
If we have a LCSSA, only modify the input value if the inval was defined
by an instruction in the loop.  If defined by something before the loop,
it is still valid.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36784 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-05 18:49:57 +00:00
Chris Lattner
a0c54f3ada make a temporary for *SI, no functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36782 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-05 18:36:36 +00:00
Chris Lattner
c42e226f08 Fix InstCombine/2007-05-04-Crash.ll and PR1384
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36775 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-05 01:59:31 +00:00
Dan Gohman
e5b01bea7b Use IntrinsicInst to test for prefetch instructions, which is ever so
slightly nicer than using CallInst with an extra check; thanks Chris.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36743 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-04 14:59:09 +00:00
Dan Gohman
2acc760165 Allow strength reduction to make use of addressing modes for the
address operand in a prefetch intrinsic.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36713 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-03 23:20:33 +00:00
Devang Patel
1997473cf7 Drop 'const'
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36662 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-03 01:11:54 +00:00
Devang Patel
3e15bf33e0 Use 'static const char' instead of 'static const int'.
Due to darwin gcc bug, one version of darwin linker coalesces
static const int, which defauts PassID based pass identification.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36652 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-02 21:39:20 +00:00
Devang Patel
794fd75c67 Do not use typeinfo to identify pass in pass manager.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36632 91177308-0d34-0410-b5e6-96231b3b80d8
2007-05-01 21:15:47 +00:00
Chris Lattner
7f3c4d49df fix a bug triggered by 403.gcc
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36527 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-28 05:27:36 +00:00
Chris Lattner
e62f021c8c Fix several latent bugs in EmitGEPOffset that didn't manifest with its
previous clients.  This fixes MallocBench/gs


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36525 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-28 04:52:43 +00:00
Chris Lattner
53c22fc7c6 uhn zap cvs
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36523 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-28 03:50:56 +00:00
Chris Lattner
9bc1464094 Implement PR1345 and Transforms/InstCombine/bitcast-gep.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36521 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-28 00:57:34 +00:00
Chris Lattner
d3e28347e5 refactor some code relating to pointer cast xforms, pulling it out of the codepath
for unrelated casts.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36511 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-27 17:44:50 +00:00
Zhou Sheng
c125c00e68 Using APInt more efficiently.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36475 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-26 16:42:07 +00:00
Devang Patel
68c01b3cf3 Mem2Reg does not need TargetData.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36444 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-25 18:32:35 +00:00
Devang Patel
41968df51e Remove unused function argument.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36441 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-25 17:15:20 +00:00
Chris Lattner
79b3bd395d If an alloca only has two types of uses: 1) reads 2) a memcpy/memmove that
copies from a constant global, then we can change the reads to read from the
global instead of from the alloca.  This eliminates the alloca and the memcpy,
and promotes secondary optimizations (because the loads are now loads from
a constant global).

This is important for a common C idiom:

void foo() {
   int A[] = {1,2,3,4,5,6,7,8,9...};
   ... only reads of A ...
}

For some reason, people forget to mark the array static or const.

This triggers on these multisource benchmarks:
JM/ldecode: block_pos, [3 x [4 x [4 x i32]]]
FreeBench/mason: m, [18 x i32], inlined 4 times
MiBench/office-stringsearch: search_strings, [1332 x i8*]
MiBench/office-stringsearch: find_strings, [1333 x i8*]
Prolangs-C++/city: dirs, [9 x i8*], inlined 4 places

and these spec benchmarks:
177.mesa: message, [8 x [32 x i8]]
186.crafty: bias_rl45, [64 x i32]
186.crafty: diag_sq, [64 x i32]
186.crafty: empty, [9 x i8]
186.crafty: xlate, [15 x i8]
186.crafty: status, [13 x i8]
186.crafty: bdinfo, [25 x i8]
445.gobmk: routines, [16 x i8*]
458.sjeng: piece_rep, [14 x i8*]
458.sjeng: t, [13 x i32], inlined 4 places.
464.h264ref: block8x8_idx, [3 x [4 x [4 x i32]]]
464.h264ref: block_pos, [3 x [4 x [4 x i32]]]
464.h264ref: j_off_tab, [12 x i32]

This implements Transforms/ScalarRepl/memcpy-from-global.ll


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36429 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-25 06:40:51 +00:00
Chris Lattner
a10b29b84b refactor the SROA code out into its own method, no functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36426 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-25 05:02:56 +00:00
Owen Anderson
ab0e4d38f0 Undo my previous changes. Since my approach to this problem is being revised,
this approach is no longer appropriate.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36421 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-25 04:18:54 +00:00
Devang Patel
75abc1ed06 Fix
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070423/048376.html


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36417 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-25 00:37:04 +00:00
Owen Anderson
3a2b58f3ad Rollback some changes that adversely affected performance. I'm currently rethinking
my approach to this, so hopefully I'll find a way to do this without making this slower.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36392 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-24 06:40:39 +00:00
Devang Patel
2a5fa18997 Fix
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070423/048333.html


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36380 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-23 22:42:03 +00:00
Owen Anderson
1b7f584fd8 Make PredicateSimplifier not use DominatorTree.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36300 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-21 07:38:12 +00:00
Owen Anderson
cdacc498e4 Fix a comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36299 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-21 07:12:44 +00:00
Owen Anderson
df07335b46 Move more passes to using ETForest instead of DominatorTree.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36271 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-20 06:27:13 +00:00
Zhou Sheng
843f0767ac Make use of ConstantInt::isZero instead of ConstantInt::isNullValue.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36261 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-19 05:39:12 +00:00
Zhou Sheng
223d65b651 Make the operations of APInt variables more efficient.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36260 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-19 05:35:00 +00:00
Owen Anderson
a65064bb3d Revert changes that caused breakage.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36255 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-18 06:46:57 +00:00
Owen Anderson
325171eed8 Switch more uses of DominatorTree over to ETForest.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36254 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-18 05:43:13 +00:00
Owen Anderson
fb4b3d117f Use ETForest instead of DominatorTree.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36252 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-18 05:25:43 +00:00
Owen Anderson
ca2563d2d8 Use ETForest instead of DominatorTree.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36249 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-18 04:55:33 +00:00
Dan Gohman
747603e39e Spell doFinalization right, so that it is a proper virtual override and
gets called.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36208 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-17 18:21:36 +00:00
Devang Patel
ac58516cdc Fix
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070416/047888.html


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36182 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-16 23:03:45 +00:00
Anton Korobeynikov
bed2946a96 Removed tabs everywhere except autogenerated & external files. Add make
target for tabs checking.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36146 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-16 18:10:23 +00:00
Owen Anderson
3dc6776b33 Remove ImmediateDominator analysis. The same information can be obtained from DomTree. A lot of code for
constructing ImmediateDominator is now folded into DomTree construction.

This is part of the ongoing work for PR217.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36063 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-15 08:47:27 +00:00
Chris Lattner
31755a024d Extend store merging to support the 'if/then' version in addition to if/then/else.
This sinks the two stores in this example into a single store in cond_next.  In this
case, it allows elimination of the load as well:

        store double 0.000000e+00, double* @s.3060
        %tmp3 = fcmp ogt double %tmp1, 5.000000e-01             ; <i1> [#uses=1]
        br i1 %tmp3, label %cond_true, label %cond_next
cond_true:              ; preds = %entry
        store double 1.000000e+00, double* @s.3060
        br label %cond_next
cond_next:              ; preds = %entry, %cond_true
        %tmp6 = load double* @s.3060            ; <double> [#uses=1]

This implements Transforms/InstCombine/store-merge.ll:test2


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36040 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-15 01:02:18 +00:00
Chris Lattner
3284d1f18a refactor some code, no functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36037 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-15 00:07:55 +00:00
Chris Lattner
e34e9a29dc fix long lines
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36031 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-14 23:32:02 +00:00
Chris Lattner
b7300fa8d2 Implement Transforms/InstCombine/vec_extract_elt.ll, transforming:
define i32 @test(float %f) {
        %tmp7 = insertelement <4 x float> undef, float %f, i32 0
        %tmp17 = bitcast <4 x float> %tmp7 to <4 x i32>
        %tmp19 = extractelement <4 x i32> %tmp17, i32 0
        ret i32 %tmp19
}

into:

define i32 @test(float %f) {
        %tmp19 = bitcast float %f to i32                ; <i32> [#uses=1]
        ret i32 %tmp19
}

On PPC, this is the difference between:

_test:
        mfspr r2, 256
        oris r3, r2, 8192
        mtspr 256, r3
        stfs f1, -16(r1)
        addi r3, r1, -16
        addi r4, r1, -32
        lvx v2, 0, r3
        stvx v2, 0, r4
        lwz r3, -32(r1)
        mtspr 256, r2
        blr

and:

_test:
        stfs f1, -4(r1)
        nop
        nop
        nop
        lwz r3, -4(r1)
        blr


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36025 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-14 23:02:14 +00:00
Chris Lattner
6987833651 Implement InstCombine/vec_demanded_elts.ll:test2. This allows us to turn
unsigned test(float f) {
 return _mm_cvtsi128_si32( (__m128i) _mm_set_ss( f*f ));
}

into:

_test:
        movss 4(%esp), %xmm0
        mulss %xmm0, %xmm0
        movd %xmm0, %eax
        ret

instead of:

_test:
        movss 4(%esp), %xmm0
        mulss %xmm0, %xmm0
        xorps %xmm1, %xmm1
        movss %xmm0, %xmm1
        movd %xmm1, %eax
        ret

GCC gets:

_test:
        subl    $28, %esp
        movss   32(%esp), %xmm0
        mulss   %xmm0, %xmm0
        xorps   %xmm1, %xmm1
        movss   %xmm0, %xmm1
        movaps  %xmm1, %xmm0
        movd    %xmm0, 12(%esp)
        movl    12(%esp), %eax
        addl    $28, %esp
        ret


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@36020 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-14 22:29:23 +00:00
Chris Lattner
6fe5541003 Implement PR1201 and test/Transforms/InstCombine/malloc-free-delete.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35981 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-14 00:20:02 +00:00
Chris Lattner
f25646bfb3 use an accessor to simplify code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35979 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-14 00:17:39 +00:00
Chris Lattner
c5494af8a9 Now that codegen prepare isn't defeating me, I can finally fix what I set
out to do! :)

This fixes a problem where LSR would insert a bunch of code into each MBB
that uses a particular subexpression (e.g. IV+base+C).  The problem is that
this code cannot be CSE'd back together if inserted into different blocks.

This patch changes LSR to attempt to insert a single copy of this code and
share it, allowing codegenprepare to duplicate the code if it can be sunk
into various addressing modes.  On CodeGen/ARM/lsr-code-insertion.ll,
for example, this gives us code like:

        add r8, r0, r5
        str r6, [r8, #+4]
..
        ble LBB1_4      @cond_next
LBB1_3: @cond_true
        str r10, [r8, #+4]
LBB1_4: @cond_next
...
LBB1_5: @cond_true55
        ldr r6, LCPI1_1
        str r6, [r8, #+4]

instead of:

        add r10, r0, r6
        str r8, [r10, #+4]
...
        ble LBB1_4      @cond_next
LBB1_3: @cond_true
        add r8, r0, r6
        str r10, [r8, #+4]
LBB1_4: @cond_next
...
LBB1_5: @cond_true55
        add r8, r0, r6
        ldr r10, LCPI1_1
        str r10, [r8, #+4]

Besides being smaller and more efficient, this makes it immediately
obvious that it is profitable to predicate LBB1_3 now :)


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35972 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-13 20:42:26 +00:00
Chris Lattner
dd77df3cbc Completely rewrite addressing-mode related sinking of code. In particular,
this fixes problems where codegenprepare would sink expressions into load/stores
that are not valid, and fixes cases where it would miss important valid ones.

This fixes several serious codesize and perf issues, particularly on targets
with complex addressing modes like arm and x86.  For example, now we compile
CodeGen/X86/isel-sink.ll to:

_test:
        movl 8(%esp), %eax
        movl 4(%esp), %ecx
        cmpl $1233, %eax
        ja LBB1_2       #F
LBB1_1: #T
        movl $4, (%ecx,%eax,4)
        movl $141, %eax
        ret
LBB1_2: #F
        movl (%ecx,%eax,4), %eax
        ret

instead of:

_test:
        movl 8(%esp), %eax
        leal (,%eax,4), %ecx
        addl 4(%esp), %ecx
        cmpl $1233, %eax
        ja LBB1_2       #F
LBB1_1: #T
        movl $4, (%ecx)
        movl $141, %eax
        ret
LBB1_2: #F
        movl (%ecx), %eax
        ret


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35970 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-13 20:30:56 +00:00
Chris Lattner
ab46275683 Fix Transforms/ScalarRepl/union-pointer.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@35906 91177308-0d34-0410-b5e6-96231b3b80d8
2007-04-11 15:45:25 +00:00