Commit Graph

90 Commits

Author SHA1 Message Date
Dan Gohman 72a13d2476 Use opt -S instead of piping bitcode output through llvm-dis.
llvm-svn: 81257
2009-09-08 22:34:10 +00:00
Dan Gohman 9737a63ed8 Change these tests to feed the assembly files to opt directly, instead
of using llvm-as, now that opt supports this.

llvm-svn: 81226
2009-09-08 16:50:01 +00:00
Dan Gohman d926b985df Create a ScalarEvolution-based AliasAnalysis implementation.
This is a simple AliasAnalysis implementation which works by making
ScalarEvolution queries. ScalarEvolution has a more complete understanding
of arithmetic than BasicAA's collection of ad-hoc checks, so it handles
some cases that BasicAA misses, for example p[i] and p[i+1] within the
same iteration of a loop.

This is currently experimental. It may be that the main use for this pass
will be to help find cases where BasicAA can be profitably extended, or
to help in the development of the overall AliasAnalysis infrastructure,
however it's also possible that it could grow up to become a directly
useful pass.

llvm-svn: 80098
2009-08-26 14:53:06 +00:00
Dan Gohman 463d3407e2 Loosen up the regex for this test so that it doesn't implicitly
depend on TargetData information.

llvm-svn: 79491
2009-08-19 23:19:36 +00:00
Dan Gohman e274526d78 Make LLVM Assembly dramatically easier to read by aligning the comments,
using formatted_raw_ostream's PadToColumn.

Before:

bb1:            ; preds = %bb
  %2 = sext i32 %i.01 to i64            ; <i64> [#uses=1]
  %3 = getelementptr double* %p, i64 %2         ; <double*> [#uses=1]
  %4 = load double* %3, align 8         ; <double> [#uses=1]
  %5 = fmul double %4, 1.100000e+00             ; <double> [#uses=1]
  %6 = sext i32 %i.01 to i64            ; <i64> [#uses=1]
  %7 = getelementptr double* %p, i64 %6         ; <double*> [#uses=1]

After:

bb1:                                        ; preds = %bb
  %2 = sext i32 %i.01 to i64                ; <i64> [#uses=1]
  %3 = getelementptr double* %p, i64 %2     ; <double*> [#uses=1]
  %4 = load double* %3, align 8             ; <double> [#uses=1]
  %5 = fmul double %4, 1.100000e+00         ; <double> [#uses=1]
  %6 = sext i32 %i.01 to i64                ; <i64> [#uses=1]
  %7 = getelementptr double* %p, i64 %6     ; <double*> [#uses=1]

Several tests required whitespace adjustments.

llvm-svn: 78816
2009-08-12 17:23:50 +00:00
Dan Gohman 9c7f808201 Change the assembly syntax for nsw, nuw, and exact, putting them
after their associated opcodes rather than before. This makes them
a little easier to read.

llvm-svn: 77194
2009-07-27 16:11:46 +00:00
Dan Gohman 534d66a426 When attempting to sign-extend an addrec by interpreting
the step value as unsigned, the start value and the addrec
itself still need to be treated as signed.

llvm-svn: 77078
2009-07-25 16:03:30 +00:00
Dan Gohman 62ef6a7f1c Teach ScalarEvolution to make use of no-overflow flags when
analyzing add recurrences.

llvm-svn: 77034
2009-07-25 01:22:26 +00:00
Dan Gohman 430f0cc544 Replace the original ad-hoc code for determining whether (v pred w) implies
(x pred y) with more thorough code that does more complete canonicalization
before resorting to range checks. This helps it find more cases where
the canonicalized expressions match.

llvm-svn: 76671
2009-07-21 23:03:19 +00:00
Dan Gohman 52e14d2272 Add a testcase for PR4569, which is now fixed.
llvm-svn: 76526
2009-07-21 00:50:52 +00:00
Dan Gohman 054d2a7837 Add testcases for PR4538, PR4537, and PR4534.
llvm-svn: 75533
2009-07-13 22:30:31 +00:00
Nick Lewycky 3292908132 When comparing constants, consider a less wide constant to be "less complex"
than a wider one, before trying to compare their contents which will crash
if their sizes are different.

llvm-svn: 74792
2009-07-04 17:24:52 +00:00
Dan Gohman 5f71a2886a Add a testcase demoing some of ScalarEvolution's new trip count logic.
llvm-svn: 74049
2009-06-24 01:22:30 +00:00
Dan Gohman 53efeb0e45 Fix a bug in the trip-count computation with And/Or. If either of the
sides is CouldNotCompute, the resulting exact count must be CouldNotCompute.

llvm-svn: 73920
2009-06-22 23:28:56 +00:00
Dan Gohman 2636693a3c Fix llvm::ComputeNumSignBits to handle pointer types
conservatively correctly, instead of aborting.

llvm-svn: 73908
2009-06-22 22:02:32 +00:00
Dan Gohman 96212b661c Teach ScalarEvolution how to analyze loops with multiple exit
blocks, and also exit blocks with multiple conditions (combined
with (bitwise) ands and ors). It's often infeasible to compute an
exact trip count in such cases, but a useful upper bound can often
be found.

llvm-svn: 73866
2009-06-22 00:31:57 +00:00
Dan Gohman 0104842ee3 Fix ScalarEvolution's backedge-taken count computations to check for
overflow when computing a integer division to round up.

Thanks to Nick Lewycky for noticing this!

llvm-svn: 73862
2009-06-21 23:46:38 +00:00
Dan Gohman eddf77123a Teach ScalarEvolution how to recognize another xor(and(x, C), C) case.
If C is a single bit and the and gets analyzed as a truncate and
zero-extend, the xor can be represnted as an add.

llvm-svn: 73664
2009-06-18 00:00:20 +00:00
Dan Gohman 432af7ace0 Add -disable-output to a bunch of tests that don't care about the output.
llvm-svn: 73633
2009-06-17 20:56:26 +00:00
Dan Gohman b50f5a46e0 Fix ScalarEvolution's Xor handling to not assume that an And
that gets recognized with a SCEVZeroExtendExpr must be an And
with a low-bits mask. With r73540, this is no longer the case.

llvm-svn: 73594
2009-06-17 01:22:39 +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
Dan Gohman 6350296efc Teach ScalarEvolution to recognize x^-1 in the case where non-demanded
bits have been stripped out by instcombine.

llvm-svn: 72010
2009-05-18 16:29:04 +00:00
Dan Gohman 8c77f1a275 Make ScalarEvolution::isLoopGuardedByCond work even when the edge
entering a loop is a non-split critical edge.

llvm-svn: 72004
2009-05-18 15:36:09 +00:00
Dan Gohman b81dd48fd2 Add nounwind to a few tests.
llvm-svn: 72002
2009-05-18 15:16:49 +00:00
Eli Friedman ebf98b0212 Allow scalar evolution to compute iteration counts for loops with a
pointer-based condition.  This fixes PR3171.

llvm-svn: 71354
2009-05-09 12:32:42 +00:00
Dan Gohman 35dc9b65ed Fix bogus overflow checks by replacing them with actual
overflow checks.

llvm-svn: 71284
2009-05-08 23:11:16 +00:00
Dan Gohman 2e55cc5a4a Fold trunc casts into add-recurrence expressions, allowing the
add-recurrence to be exposed. Add a new SCEV folding rule to
help simplify expressions in the presence of these extra truncs.

llvm-svn: 71264
2009-05-08 21:03:19 +00:00
Dan Gohman 7227bc88f0 When printing a SCEVUnknown with pointer type, don't print an
artificial "ptrtoint", as it tends to clutter up complicated
expressions. The cast operators now print both source and
destination types, which is usually sufficient.

llvm-svn: 70554
2009-05-01 17:02:22 +00:00
Dan Gohman 2b8da35f9d Extend ScalarEvolution's getBackedgeTakenCount to be able to
compute an upper-bound value for the trip count, in addition to
the actual trip count. Use this to allow getZeroExtendExpr and
getSignExtendExpr to fold casts in more cases.

This may eventually morph into a more general value-range
analysis capability; there are certainly plenty of places where
more complete value-range information would allow more folding.

llvm-svn: 70509
2009-04-30 20:47:05 +00:00
Dan Gohman 494dac3f84 Generalize the cast-of-addrec folding to handle folding of SCEVs like
(sext i8 {-128,+,1} to i64) to i64 {-128,+,1}, where the iteration
crosses from negative to positive, but is still safe if the trip
count is within range.

llvm-svn: 70421
2009-04-29 22:28:28 +00:00
Dan Gohman d9775a3be1 Fix this test to match the new output from scalar-evolution.
llvm-svn: 70410
2009-04-29 21:06:20 +00:00
Dan Gohman d9b11b2ef4 Include the source type in SCEV cast expression debug output, and
print sext, zext, and trunc, instead of signextend, zeroextend,
and truncate, respectively, for consistency with the main IR.

llvm-svn: 70405
2009-04-29 20:27:52 +00:00
Dan Gohman 807dff7486 Fix a grammaro in a comment.
llvm-svn: 70331
2009-04-28 21:54:23 +00:00
Nick Lewycky b4d9f7a9b3 Simplify trunc(extend(x)) in SCEVs, just for completeness. Also fix some odd
whitespace in the same file.

llvm-svn: 69870
2009-04-23 05:15:08 +00:00
Owen Anderson 7d82244be7 Testcase for PR3909.
llvm-svn: 69868
2009-04-23 04:33:42 +00:00
Dan Gohman e14efcc9f4 When turning (ashr(shl(x, n), n)) into sext(trunc(x)), the width of the
type to truncate to should be the number of bits of the value that are
preserved, not the number that are clobbered with sign-extension.
This fixes regressions in ldecod.

llvm-svn: 69704
2009-04-21 20:18:36 +00:00
Dan Gohman 0bddac16a8 Rename ScalarEvolution's getIterationCount to getBackedgeTakenCount,
to more accurately describe what it does. Expand its doxygen comment
to describe what the backedge-taken count is and how it differs
from the actual iteration count of the loop. Adjust names and
comments in associated code accordingly.

llvm-svn: 65382
2009-02-24 18:55:53 +00:00
Nick Lewycky 52348300a4 Wind SCEV back in time, to Nov 18th. This 'fixes' PR3275, PR3294, PR3295,
PR3296 and PR3302.

llvm-svn: 62160
2009-01-13 09:18:58 +00:00
Nick Lewycky 380292a51a Don't try to analyze this "backward" case. This is overly conservative
pending a correct solution.

llvm-svn: 61589
2009-01-02 18:54:17 +00:00
Nick Lewycky 69c9aa4ce5 Generalize support for analyzing loops to include SLE/SGE loop exit conditions
and support for non-unit strides with signed exit conditions.

llvm-svn: 61082
2008-12-16 08:30:01 +00:00
Nick Lewycky 729bf137a8 Revert my re-instated reverted commit, fixes the bootstrap build on x86-64 linux.
llvm-svn: 60951
2008-12-12 17:09:07 +00:00
Nick Lewycky 6a344e097c Sneaky, sneaky: move the -1 to the outside of the SMax. Reinstate the
optimization of SGE/SLE with unit stride, now that it works properly.

llvm-svn: 60881
2008-12-11 17:40:14 +00:00
Evan Cheng 058522f1da xfail this for now.
llvm-svn: 60777
2008-12-09 18:43:00 +00:00
Nick Lewycky f545749f2b It's easy to handle SLE/SGE when the loop has a unit stride.
llvm-svn: 60748
2008-12-09 07:25:04 +00:00
Nick Lewycky 1c451ae43e Add a utility function that detects whether a loop is guaranteed to be finite.
Use it to safely handle less-than-or-equals-to exit conditions in loops. These
also occur when the loop exit branch is exit on true because SCEV inverses the
icmp predicate.

Use it again to handle non-zero strides, but only with an unsigned comparison
in the exit condition.

llvm-svn: 59528
2008-11-18 15:10:54 +00:00
Nick Lewycky 625c6f79b2 Don't brute-force analyze cubic or higher polynomials.
If this patch causes a performance regression for anyone, please let me know,
and it can be fixed in a different way with much more effort.

llvm-svn: 59384
2008-11-16 04:14:25 +00:00
Nick Lewycky 7b14e20a5e Don't crash analyzing certain quadratics (addrec of {X,+,Y,+,1}).
We're still waiting on code that actually analyzes them properly.

llvm-svn: 58592
2008-11-03 02:43:49 +00:00
Dan Gohman dc5f5cbe59 Finally re-apply r46959. This is made feasible by the combination
of r56230, r56232, and r56246.

llvm-svn: 56247
2008-09-16 18:52:57 +00:00
Dan Gohman 162568842e Fix spacing in the grep line for this test, following the recent
SCEV-whitespace changes.

llvm-svn: 56234
2008-09-16 01:37:08 +00:00
Dan Gohman f9081a2cd5 Teach ScalarEvolution to consider loop preheaders in the search for
an if statement that guards a loop, to allow indvars to avoid smax
operations in more situations.

llvm-svn: 56232
2008-09-15 22:18:04 +00:00