infrastructure.
This was essentially work toward PGO based on a design that had several
flaws, partially dating from a time when LLVM had a different
architecture, and with an effort to modernize it abandoned without being
completed. Since then, it has bitrotted for several years further. The
result is nearly unusable, and isn't helping any of the modern PGO
efforts. Instead, it is getting in the way, adding confusion about PGO
in LLVM and distracting everyone with maintenance on essentially dead
code. Removing it paves the way for modern efforts around PGO.
Among other effects, this removes the last of the runtime libraries from
LLVM. Those are being developed in the separate 'compiler-rt' project
now, with somewhat different licensing specifically more approriate for
runtimes.
llvm-svn: 191835
Remove the command line argument "struct-path-tbaa" since we should not depend
on command line argument to decide which format the IR file is using. Instead,
we check the first operand of the tbaa tag node, if it is a MDNode, we treat
it as struct-path aware TBAA format, otherwise, we treat it as scalar TBAA
format.
When clang starts to use struct-path aware TBAA format no matter whether
struct-path-tbaa is no, and we can auto-upgrade existing bc files, the support
for scalar TBAA format can be dropped.
Existing testing cases are updated to use the struct-path aware TBAA format.
llvm-svn: 191538
This code isn't ready to deal with allocation functions where the return is not
the allocated pointer. The checks below will reject posix_memalign anyways.
llvm-svn: 191319
This is safe per C++11 18.6.1.1p3: [operator new returns] a non-null pointer to
suitably aligned storage (3.7.4), or else throw a bad_alloc exception. This
requirement is binding on a replacement version of this function.
Brings us a tiny bit closer to eliminating more vector push_backs.
llvm-svn: 191310
Overflow doesn't affect the correctness of equalities. Computing this is cheap,
we just reuse the computation for the inbounds case and try to peel of more
non-inbounds GEPs. This pattern is unlikely to ever appear in code generated by
Clang, but SCEV occasionally produces it.
llvm-svn: 191200
Upcoming SLP vectorization improvements will want to be able to estimate costs
of horizontal reductions. Add infrastructure to support this.
We model reductions as a series of (shufflevector,add) tuples ultimately
followed by an extractelement. For example, for an add-reduction of <4 x float>
we could generate the following sequence:
(v0, v1, v2, v3)
\ \ / /
\ \ /
+ +
(v0+v2, v1+v3, undef, undef)
\ /
((v0+v2) + (v1+v3), undef, undef)
%rdx.shuf = shufflevector <4 x float> %rdx, <4 x float> undef,
<4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
%bin.rdx = fadd <4 x float> %rdx, %rdx.shuf
%rdx.shuf7 = shufflevector <4 x float> %bin.rdx, <4 x float> undef,
<4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
%bin.rdx8 = fadd <4 x float> %bin.rdx, %rdx.shuf7
%r = extractelement <4 x float> %bin.rdx8, i32 0
This commit adds a cost model interface "getReductionCost(Opcode, Ty, Pairwise)"
that will allow clients to ask for the cost of such a reduction (as backends
might generate more efficient code than the cost of the individual instructions
summed up). This interface is excercised by the CostModel analysis pass which
looks for reduction patterns like the one above - starting at extractelements -
and if it sees a matching sequence will call the cost model interface.
We will also support a second form of pairwise reduction that is well supported
on common architectures (haddps, vpadd, faddp).
(v0, v1, v2, v3)
\ / \ /
(v0+v1, v2+v3, undef, undef)
\ /
((v0+v1)+(v2+v3), undef, undef, undef)
%rdx.shuf.0.0 = shufflevector <4 x float> %rdx, <4 x float> undef,
<4 x i32> <i32 0, i32 2 , i32 undef, i32 undef>
%rdx.shuf.0.1 = shufflevector <4 x float> %rdx, <4 x float> undef,
<4 x i32> <i32 1, i32 3, i32 undef, i32 undef>
%bin.rdx.0 = fadd <4 x float> %rdx.shuf.0.0, %rdx.shuf.0.1
%rdx.shuf.1.0 = shufflevector <4 x float> %bin.rdx.0, <4 x float> undef,
<4 x i32> <i32 0, i32 undef, i32 undef, i32 undef>
%rdx.shuf.1.1 = shufflevector <4 x float> %bin.rdx.0, <4 x float> undef,
<4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
%bin.rdx.1 = fadd <4 x float> %rdx.shuf.1.0, %rdx.shuf.1.1
%r = extractelement <4 x float> %bin.rdx.1, i32 0
llvm-svn: 190876
Allow targets to customize the default behavior of the generic loop unrolling
transformation. This will be used by the PowerPC backend when targeting the A2
core (which is in-order with a deep pipeline), and using more aggressive
defaults is important.
llvm-svn: 190542
instead of having its own implementation.
The implementation of isTBAAVtableAccess is in TypeBasedAliasAnalysis.cpp
since it is related to the format of TBAA metadata.
The path for struct-path tbaa will be exercised by
test/Instrumentation/ThreadSanitizer/read_from_global.ll, vptr_read.ll, and
vptr_update.ll when struct-path tbaa is on by default.
llvm-svn: 190216
Revert unintentional commit (of an unreviewed change).
Original commit message:
Add getUnrollingPreferences to TTI
Allow targets to customize the default behavior of the generic loop unrolling
transformation. This will be used by the PowerPC backend when targeting the A2
core (which is in-order with a deep pipeline), and using more aggressive
defaults is important.
llvm-svn: 189566
Allow targets to customize the default behavior of the generic loop unrolling
transformation. This will be used by the PowerPC backend when targeting the A2
core (which is in-order with a deep pipeline), and using more aggressive
defaults is important.
llvm-svn: 189565
...so that it can be used for z too. Most of the code is the same.
The only real change is to use TargetTransformInfo to test when a sqrt
instruction is available.
The pass is opt-in because at the moment it only handles sqrt.
llvm-svn: 189097
This fixes SCEVExpander so that it does not create multiple distinct induction
variables for duplicate PHI entries. Specifically, given some code like this:
do.body6: ; preds = %do.body6, %do.body6, %if.then5
%end.0 = phi i8* [ undef, %if.then5 ], [ %incdec.ptr, %do.body6 ], [ %incdec.ptr, %do.body6 ]
...
Note that it is legal to have multiple entries for a basic block so long as the
associated value is the same. So the above input is okay, but expanding an
AddRec in this loop could produce code like this:
do.body6: ; preds = %do.body6, %do.body6, %if.then5
%indvar = phi i64 [ %indvar.next, %do.body6 ], [ %indvar.next1, %do.body6 ], [ 0, %if.then5 ]
%end.0 = phi i8* [ undef, %if.then5 ], [ %incdec.ptr, %do.body6 ], [ %incdec.ptr, %do.body6 ]
...
%indvar.next = add i64 %indvar, 1
%indvar.next1 = add i64 %indvar, 1
And this is not legal because there are two PHI entries for %do.body6 each with
a distinct value.
Unfortunately, I don't have an in-tree test case.
llvm-svn: 188614
to find loops if the From and To instructions were in the same block.
Refactor the code a little now that we need to fill to start the CFG-walking
algorithm with more than one starting basic block sometimes.
Special thanks to Andrew Trick for catching an error in my understanding of
natural loops in code review.
llvm-svn: 188236
All libm floating-point rounding functions, except for round(), had their own
ISD nodes. Recent PowerPC cores have an instruction for round(), and so here I'm
adding ISD::FROUND so that round() can be custom lowered as well.
For the most part, this is straightforward. I've added an intrinsic
and a matching ISD node just like those for nearbyint() and friends. The
SelectionDAG pattern I've named frnd (because ISD::FP_ROUND has already claimed
fround).
This will be used by the PowerPC backend in a follow-up commit.
llvm-svn: 187926
This fix is very lightweight. The same fix already existed for AddRec
but was missing for NAry expressions.
This is obviously an improvement and I'm unsure how to test compile
time problems.
Patch by Xiaoyi Guo!
llvm-svn: 187475
Call into ComputeMaskedBits to figure out which bits are set on both add
operands and determine if the value is a power-of-two-or-zero or not.
llvm-svn: 187445
Adds unit tests for it too.
Split BasicBlockUtils into an analysis-half and a transforms-half, and put the
analysis bits into a new Analysis/CFG.{h,cpp}. Promote isPotentiallyReachable
into llvm::isPotentiallyReachable and move it into Analysis/CFG.
llvm-svn: 187283
Merge consecutive if-regions if they contain identical statements.
Both transformations reduce number of branches. The transformation
is guarded by a target-hook, and is currently enabled only for +R600,
but the correctness has been tested on X86 target using a variety of
CPU benchmarks.
Patch by: Mei Ye
llvm-svn: 187278
The great thing about the SCEVAddRec No-Wrap flag (unlike nsw/nuw) is
that is can be preserved while normalizing (reassociating and
factoring).
The bad thing is that is can't be tranfered back to IR, which is one
of the reasons I don't like the concept of SCEVExpander.
Sorry, I can't think of a direct way to test this, which is why these
were FIXMEs for so long. I just think it's a good time to finally
clean it up.
llvm-svn: 186273
Address calculation for gather/scather in vectorized code can incur a
significant cost making vectorization unbeneficial. Add infrastructure to add
cost.
Tests and cost model for targets will be in follow-up commits.
radar://14351991
llvm-svn: 186187
ScalarEvolution::getSignedRange uses ComputeNumSignBits from ValueTracking on
ashr instructions. ComputeNumSignBits can return zero, but this case was not
handled correctly by the code in getSignedRange which was calling:
APInt::getSignedMinValue(BitWidth).ashr(NS - 1)
with NS = 0, resulting in an assertion failure in APInt::ashr.
Now, we just return the conservative result (as with NS == 1).
Another bug found by llvm-stress.
llvm-svn: 185955
(add nsw x, (and x, y)) isn't a power of two if x is zero, it's zero
(add nsw x, (xor x, y)) isn't a power of two if y has bits set that aren't set in x
llvm-svn: 185954
The symptom is seg-fault, and the root cause is that a SCEV contains a SCEVUnknown
which has null-pointer to a llvm::Value.
This is how the problem take place:
===================================
1). In the pristine input IR, there are two relevant instrutions Op1 and Op2,
Op1's corresponding SCEV (denoted as SCEV(op1)) is a SCEVUnknown, and
SCEV(Op2) contains SCEV(Op1). None of these instructions are dead.
Op1 : V1 = ...
...
Op2 : V2 = ... // directly or indirectly (data-flow) depends on Op1
2) Optimizer (LSR in my case) generates an instruction holding the equivalent
value of Op1, making Op1 dead.
Op1': V1' = ...
Op1: V1 = ... ; now dead)
Op2 : V2 = ... //Now deps on Op1', but the SCEV(Op2) still contains SCEV(Op1)
3) Op1 is deleted, and call-back function is called to reset
SCEV(Op1) to indicate it is invalid. However, SCEV(Op2) is not
invalidated as well.
4) Following pass get the cached, invalid SCEV(Op2), and try to manipulate it,
and cause segfault.
The fix:
========
It seems there is no clean yet inexpensive fix. I write to dev-list
soliciting good solution, unforunately no ack. So, I decide to fix this
problem in a brute-force way:
When ScalarEvolution::getSCEV is called, check if the cached SCEV
contains a invalid SCEVUnknow, if yes, remove the cached SCEV, and
re-evaluate the SCEV from scratch.
I compile buch of big *.c and *.cpp, fortunately, I don't see any increase
in compile time.
Misc:
=====
The reduced test-case has 2357 lines of code+other-stuff, too big to commit.
rdar://14283433
llvm-svn: 185843
The Builtin attribute is an attribute that can be placed on function call site that signal that even though a function is declared as being a builtin,
rdar://problem/13727199
llvm-svn: 185049
This is a band-aid to fix the most severe regressions we're seeing from basing
spill decisions on block frequencies, until we have a better solution.
llvm-svn: 184835
Fixes rdar:14036816, PR16130.
There is an opportunity to compute precise trip counts for 'or'
expressions and multi-exit loops.
rdar:14038809: Optimize trip count computation for multi-exit loops.
To do this we need to record the fact that ExitLimit assumes NSW. When
it does not we can safely assume that the loop trip count is the
minimum ExitLimt across all subexpressions and loop exits.
llvm-svn: 183060
Account for the cost of scaling factor in Loop Strength Reduce when rating the
formulae. This uses a target hook.
The default implementation of the hook is: if the addressing mode is legal, the
scaling factor is free.
<rdar://problem/13806271>
llvm-svn: 183045
Fixes PR16130 - clang produces incorrect code with loop/expression at -O2.
This is a 2+ year old bug that's now holding up the release. It's a
case where we knowingly made aggressive assumptions about undefined
behavior. These assumptions are wrong when SCEV is computing a
subexpression that does not directly control the branch. With this
fix, we avoid making assumptions in those cases but still optimize the
common case. SCEV's trip count computation for exits controlled by
'or' expressions is now analagous to the trip count computation for
loops with multiple exits. I had already fixed the multiple exit case
to be conservative.
llvm-svn: 182989
- llvm.loop.parallel metadata has been renamed to llvm.loop to be more generic
by making the root of additional loop metadata.
- Loop::isAnnotatedParallel now looks for llvm.loop and associated
llvm.mem.parallel_loop_access
- document llvm.loop and update llvm.mem.parallel_loop_access
- add support for llvm.vectorizer.width and llvm.vectorizer.unroll
- document llvm.vectorizer.* metadata
- add utility class LoopVectorizerHints for getting/setting loop metadata
- use llvm.vectorizer.width=1 to indicate already vectorized instead of
already_vectorized
- update existing tests that used llvm.loop.parallel and
llvm.vectorizer.already_vectorized
Reviewed by: Nadav Rotem
llvm-svn: 182802
Other than recognizing the attribute, the patch does little else.
It changes the branch probability analyzer so that edges into
blocks postdominated by a cold function are given low weight.
Added analysis and code generation tests. Added documentation for the
new attribute.
llvm-svn: 182638
BitVector/SmallBitVector::reference::operator bool remain implicit since
they model more exactly a bool, rather than something else that can be
boolean tested.
The most common (non-buggy) case are where such objects are used as
return expressions in bool-returning functions or as boolean function
arguments. In those cases I've used (& added if necessary) a named
function to provide the equivalent (or sometimes negative, depending on
convenient wording) test.
One behavior change (YAMLParser) was made, though no test case is
included as I'm not sure how to reach that code path. Essentially any
comparison of llvm::yaml::document_iterators would be invalid if neither
iterator was at the end.
This helped uncover a couple of bugs in Clang - test cases provided for
those in a separate commit along with similar changes to `operator bool`
instances in Clang.
llvm-svn: 181868
the things, and renames it to CBindingWrapping.h. I also moved
CBindingWrapping.h into Support/.
This new file just contains the macros for defining different wrap/unwrap
methods.
The calls to those macros, as well as any custom wrap/unwrap definitions
(like for array of Values for example), are put into corresponding C++
headers.
Doing this required some #include surgery, since some .cpp files relied
on the fact that including Wrap.h implicitly caused the inclusion of a
bunch of other things.
This also now means that the C++ headers will include their corresponding
C API headers; for example Value.h must include llvm-c/Core.h. I think
this is harmless, since the C API headers contain just external function
declarations and some C types, so I don't believe there should be any
nasty dependency issues here.
llvm-svn: 180881
We switch the order of offset and field type to make TBAAStructType node
(name, parent node, offset) similar to scalar TBAA node (name, parent node).
TypeIsImmutable is added to TBAAStructTag node.
llvm-svn: 180654
The tag is of type TBAANode when flag EnableStructPathTBAA is off.
Move implementation of MDNode::getMostGenericTBAA to TypeBasedAliasAnalysis.cpp
since it depends on how to interprete the MDNodes for scalar TBAA and
struct-path aware TBAA.
llvm-svn: 180068
PR15000 has a testcase where the time to compile was bordering on 30s. When I
dropped the limit value to 100, it became a much more managable 6s. The compile
time seems to increase in a roughly linear fashion based on increasing the limit
value. (See the runtimes below.)
So, let's lower the limit to 100 so that they can get a more reasonable compile
time.
Limit Value Time
----------- ----
10 0.9744s
20 1.8035s
30 2.3618s
40 2.9814s
50 3.6988s
60 4.5486s
70 4.9314s
80 5.8012s
90 6.4246s
100 7.0852s
110 7.6634s
120 8.3553s
130 9.0552s
140 9.6820s
150 9.8804s
160 10.8901s
170 10.9855s
180 12.0114s
190 12.6816s
200 13.2754s
210 13.9942s
220 13.8097s
230 14.3272s
240 15.7753s
250 15.6673s
260 16.0541s
270 16.7625s
280 17.3823s
290 18.8213s
300 18.6120s
310 20.0333s
320 19.5165s
330 20.2505s
340 20.7068s
350 21.1833s
360 22.9216s
370 22.2152s
380 23.9390s
390 23.4609s
400 24.0426s
410 24.6410s
420 26.5208s
430 27.7155s
440 26.4142s
450 28.5646s
460 27.3494s
470 29.7255s
480 29.4646s
490 30.5001s
llvm-svn: 179713
This is basically the same fix in three different places. We use a set to avoid
walking the whole tree of a big ConstantExprs multiple times.
For example: (select cmp, (add big_expr 1), (add big_expr 2))
We don't want to visit big_expr twice here, it may consist of thousands of
nodes.
The testcase exercises this by creating an insanely large ConstantExprs out of
a loop. It's questionable if the optimizer should ever create those, but this
can be triggered with real C code. Fixes PR15714.
llvm-svn: 179458
Chandler Carruth