294 Commits

Author SHA1 Message Date
Wojciech Matyjewicz
1d2c27b23e Fix PR2002. Suppose n is the initial value for the induction
variable (with step 1) and m is its final value. Then, the correct trip 
count is SMAX(m,n)-n. Previously, we used SMAX(0,m-n), but m-n may 
overflow and can't in general be interpreted as signed.

Patch by Nick Lewycky.

llvm-svn: 47007
2008-02-12 15:09:36 +00:00
Wojciech Matyjewicz
adae053b53 If the LHS of the comparison is a loop-invariant we also want to move it
to the RHS. This simple change allows to compute loop iteration count 
for loops with condition similar to the one in the testcase (which seems 
to be quite common).

llvm-svn: 46959
2008-02-11 18:37:34 +00:00
Wojciech Matyjewicz
d2d9764cc8 Fix PR1798 - an error in the evaluation of SCEVAddRecExpr at an
arbitrary iteration.

The patch:
1) changes SCEVSDivExpr into SCEVUDivExpr,
2) replaces PartialFact() function with BinomialCoefficient(); the 
computations (essentially, the division) in BinomialCoefficient() are 
performed with the apprioprate bitwidth necessary to avoid overflow; 
unsigned division is used instead of the signed one.

Computations in BinomialCoefficient() require support from the code 
generator for APInts. Currently, we use a hack rounding up the 
neccessary bitwidth to the nearest power of 2. The hack is easy to turn 
off in future.

One remaining issue: we assume the divisor of the binomial coefficient 
formula can be computed accurately using 16 bits. It means we can handle 
AddRecs of length up to 9. In future, we should use APInts to evaluate 
the divisor.

Thanks to Nicholas for cooperation!

llvm-svn: 46955
2008-02-11 11:03:14 +00:00
Dan Gohman
a65951fef0 Avoid unnecessarily casting away const, fixing a FIXME.
llvm-svn: 46591
2008-01-31 01:05:10 +00:00
Nick Lewycky
d1200b00dc Don't be rude, emit debugging info where asked to.
llvm-svn: 45485
2008-01-02 02:49:20 +00:00
Chris Lattner
f3ebc3f3d2 Remove attribution from file headers, per discussion on llvmdev.
llvm-svn: 45418
2007-12-29 20:36:04 +00:00
Chris Lattner
d2265b45ae Fix PR1850 by removing an unsafe transformation from VMCore/ConstantFold.cpp.
Reimplement the xform in Analysis/ConstantFolding.cpp where we can use
targetdata to validate that it is safe.  While I'm in there, fix some const
correctness issues and generalize the interface to the "operand folder".

llvm-svn: 44817
2007-12-10 22:53:04 +00:00
Nick Lewycky
cdb7e54ca7 Add new SCEV, SCEVSMax. This allows LLVM to analyze do-while loops.
llvm-svn: 44319
2007-11-25 22:41:31 +00:00
Chris Lattner
69ec1ec86a simplify some code.
llvm-svn: 44295
2007-11-23 22:36:49 +00:00
Chris Lattner
a8fbde3f78 Fix a bug where we'd try to find a scev value for a bitcast operand,
even though the bitcast operand did not have integer type.  This fixes
PR1814.

llvm-svn: 44286
2007-11-23 08:46:22 +00:00
Nick Lewycky
3783b46f9e Instead of calculating constant factors, calculate the number of trailing
bits. Patch from Wojciech Matyjewicz.

llvm-svn: 44268
2007-11-22 07:59:40 +00:00
Nick Lewycky
74a26e3750 Small cleanup. Use APInt::getHighBitsSet method instead of shift left.
"setcc" -> "icmp op" in comments. No functionality change.

llvm-svn: 44249
2007-11-20 08:44:50 +00:00
Nick Lewycky
5b18bd3368 Be more careful when transforming | to +. Patch from Wojciech Matyjewicz.
llvm-svn: 44248
2007-11-20 08:24:44 +00:00
Anton Korobeynikov
6a7ddfdb8f Reverted r44163 per request
llvm-svn: 44177
2007-11-15 18:33:16 +00:00
Nick Lewycky
fbb24817cc Fix handling of overflow in loop calculation by adding new UDiv SCEV. This SCEV
is disabled in the sense that it will refuse to create one from a UDiv
instruction, until the code is better tested.

llvm-svn: 44163
2007-11-15 06:30:50 +00:00
Dan Gohman
a37eaf2bf9 Move the SCEV object factors from being static members of the individual
SCEV subclasses to being non-static member functions of the ScalarEvolution
class.

llvm-svn: 43224
2007-10-22 18:31:58 +00:00
Nick Lewycky
3934961878 Build the correct range for loops with unusual bounds. Fix from Jay Foad.
llvm-svn: 42394
2007-09-27 14:12:54 +00:00
Dale Johannesen
bed9dc423c Next round of APFloat changes.
Use APFloat in UpgradeParser and AsmParser.
Change all references to ConstantFP to use the
APFloat interface rather than double.  Remove
the ConstantFP double interfaces.
Use APFloat functions for constant folding arithmetic
and comparisons.
(There are still way too many places APFloat is
just a wrapper around host float/double, but we're
getting there.)

llvm-svn: 41747
2007-09-06 18:13:44 +00:00
Devang Patel
b5933bbbd5 Use SmallVector instead of std::vector.
llvm-svn: 41207
2007-08-21 00:31:24 +00:00
Nick Lewycky
96606cec20 Let scalar-evolution analyze loops with an unsigned comparison for the exit
condition. Fixes 1597.

llvm-svn: 40867
2007-08-06 19:21:00 +00:00
Nick Lewycky
b9819f3a8b Don't assume it's safe to transform a loop just because it's dominated by any
comparison. Fixes bug 1598.

llvm-svn: 40866
2007-08-06 18:33:46 +00:00
Nick Lewycky
5246026c8c Handle decrementing loops properly. Fixes PR1533.
Always pass the constant as the second parameter to HowManyLessThans.

Remove obsolete "isSigned" parameter.

llvm-svn: 39893
2007-07-16 02:08:00 +00:00
Dan Gohman
0a76e7f678 Move the APInt form of SCEVUnknown::getIntegerSCEV to SCEVConstant::get, and
use SCEVConstant::get instead of SCEVUnknown::get when constructing a SCEV
for a ConstantInt.

llvm-svn: 38457
2007-07-09 15:25:17 +00:00
Dan Gohman
32f53bbd85 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.

llvm-svn: 37657
2007-06-19 14:28:31 +00:00
Dan Gohman
eed125f315 In SCEVAddExpr::get, skip over any cast operands before looking for nested
add operands after constant operands. The recent change to recognize
sign-extend expressions caused this to be exposed more often.

llvm-svn: 37628
2007-06-18 19:30:09 +00:00
Dan Gohman
cb9e09ad57 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.

llvm-svn: 37591
2007-06-15 14:38:12 +00:00
Nick Lewycky
df543f4b84 Optimize this test. Firstly, only Instructions may use other Instructions.
Secondly, checking whether removal succeeded tells you whether it was in
the map to begin with.

llvm-svn: 37469
2007-06-06 11:26:20 +00:00
Nick Lewycky
3e84212897 Fix PR1487 and Transforms/IndVar/2007-06-06-DeleteDanglesPtr.ll
llvm-svn: 37459
2007-06-06 04:12:20 +00:00
Devang Patel
8c78a0bff0 Drop 'const'
llvm-svn: 36662
2007-05-03 01:11:54 +00:00
Devang Patel
e95c6ad802 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.

llvm-svn: 36652
2007-05-02 21:39:20 +00:00
Lauro Ramos Venancio
41223586a2 Fix build error.
llvm-svn: 36648
2007-05-02 20:37:47 +00:00
Devang Patel
09f162ca6a Do not use typeinfo to identify pass in pass manager.
llvm-svn: 36632
2007-05-01 21:15:47 +00:00
Reid Spencer
a3cfb8a683 Revert last patch. It was already fixed.
llvm-svn: 36102
2007-04-16 02:24:41 +00:00
Reid Spencer
8be22e4e04 For PR1336:
Fix a div-by-zero bug noticed by APInt. This fixes:
test/Transforms/IndVarsSimplify/exit_value_tests.llx

llvm-svn: 36099
2007-04-16 01:48:37 +00:00
Chris Lattner
cad61e81c1 Fix a nasty bug introduced when apint'ified. This fixes
Transforms/IndVarsSimplify/exit_value_tests.llx

llvm-svn: 36081
2007-04-15 19:52:49 +00:00
Zhou Sheng
2852d99a48 Eliminate unnecessary APInt construction.
llvm-svn: 35740
2007-04-07 17:48:27 +00:00
Zhou Sheng
c0297896aa Make APInt variables do the computation stuffs instead of
ConstantExpr::getXX if possible.

llvm-svn: 35738
2007-04-07 17:40:57 +00:00
Zhou Sheng
1459c73481 Eliminate unnecessary zext/trunc stuffs.
llvm-svn: 35737
2007-04-07 17:12:38 +00:00
Chris Lattner
a3e0bb4ebb Treat xor of signbit like an add.
llvm-svn: 35586
2007-04-02 05:41:38 +00:00
Reid Spencer
69904f939a Guard further against APInt operations with operands of unequal bit width.
llvm-svn: 34897
2007-03-04 01:25:35 +00:00
Reid Spencer
e1336b4086 Fix an unequal bitwidth issue.
llvm-svn: 34831
2007-03-02 02:59:25 +00:00
Reid Spencer
2e54a15943 Prefer non-virtual calls to ConstantInt::isZero over virtual calls to
Constant::isNullValue() in situations where it is possible.

llvm-svn: 34821
2007-03-02 00:28:52 +00:00
Reid Spencer
caace8cffd Make it possible to create an SCEVUnknown from an APInt as well as an int.
llvm-svn: 34816
2007-03-01 22:28:51 +00:00
Reid Spencer
fad3f247e4 Construct ConstantInt with simpler constructor.
llvm-svn: 34795
2007-03-01 19:32:33 +00:00
Reid Spencer
6ba8edc789 Fix last night's 445.gobmk breakage which was caused by comparison of
APInt's of unequal bitwidth.

llvm-svn: 34790
2007-03-01 17:17:21 +00:00
Reid Spencer
6a44033465 Remove the "isSigned" parameters from ConstantRange. It turns out they
are not needed as the results are the same with or without it.

Patch by Nicholas Lewycky.

llvm-svn: 34782
2007-03-01 07:54:15 +00:00
Reid Spencer
983e3b37de APIntify various computations in ScalarEvolution
llvm-svn: 34780
2007-03-01 07:25:48 +00:00
Reid Spencer
b0170c4259 For PR1205:
Make GetConstantFactor compute its result using an APInt.

llvm-svn: 34765
2007-02-28 23:31:17 +00:00
Reid Spencer
d373b9dc59 For PR1205:
Adjust to changes in ConstantRange interface.

llvm-svn: 34762
2007-02-28 22:03:51 +00:00
Reid Spencer
3a7e9d8e75 For PR1205:
Remove ConstantInt from ConstantRange interface and adjust its users to
compensate.

llvm-svn: 34758
2007-02-28 19:57:34 +00:00