check more intelligent. This speeds up mem2reg from 5.29s to
0.79s on a synthetic testcase with tons of predecessors and
phi nodes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@46767 91177308-0d34-0410-b5e6-96231b3b80d8
inlining a function if we know that the function does not write
to *any* memory. This implements test/Transforms/Inline/byval2.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45912 91177308-0d34-0410-b5e6-96231b3b80d8
could theoretically introduce a trap, but is also a performance issue.
This speeds up ptrdist/ks by 8%.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45533 91177308-0d34-0410-b5e6-96231b3b80d8
define void @f() {
...
call i32 @g()
...
}
define void @g() {
...
}
The hazards are:
- @f and @g have GC, but they differ GC. Inlining is invalid. This
may never occur.
- @f has no GC, but @g does. g's GC must be propagated to @f.
The other scenarios are safe:
- @f and @g have the same GC.
- @f and @g have no GC.
- @g has no GC.
This patch adds inliner checks for the former two scenarios.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45351 91177308-0d34-0410-b5e6-96231b3b80d8
calls 'nounwind'. It is important for correct C++
exception handling that nounwind markings do not get
lost, so this transformation is actually needed for
correctness.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45218 91177308-0d34-0410-b5e6-96231b3b80d8
how to lower them (with no attempt made to be
efficient, since they should only occur for
unoptimized code).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45108 91177308-0d34-0410-b5e6-96231b3b80d8
calls. Remove special casing of inline asm from the
inliner. There is a potential problem: the verifier
rejects invokes of inline asm (not sure why). If an
asm call is not marked "nounwind" in some .ll, and
instcombine is not run, but the inliner is run, then
an illegal module will be created. This is bad but
I'm not sure what the best approach is. I'm tempted
to remove the check in the verifier...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45073 91177308-0d34-0410-b5e6-96231b3b80d8
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".
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@44817 91177308-0d34-0410-b5e6-96231b3b80d8
methods are new to Function:
bool hasCollector() const;
const std::string &getCollector() const;
void setCollector(const std::string &);
void clearCollector();
The assembly representation is as such:
define void @f() gc "shadow-stack" { ...
The implementation uses an on-the-side table to map Functions to
collector names, such that there is no overhead. A StringPool is
further used to unique collector names, which are extremely
likely to be unique per process.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@44769 91177308-0d34-0410-b5e6-96231b3b80d8
throw exceptions", just mark intrinsics with the nounwind
attribute. Likewise, mark intrinsics as readnone/readonly
and get rid of special aliasing logic (which didn't use
anything more than this anyway).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@44544 91177308-0d34-0410-b5e6-96231b3b80d8
the function type, instead they belong to functions
and function calls. This is an updated and slightly
corrected version of Reid Spencer's original patch.
The only known problem is that auto-upgrading of
bitcode files doesn't seem to work properly (see
test/Bitcode/AutoUpgradeIntrinsics.ll). Hopefully
a bitcode guru (who might that be? :) ) will fix it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@44359 91177308-0d34-0410-b5e6-96231b3b80d8
The meaning of getTypeSize was not clear - clarifying it is important
now that we have x86 long double and arbitrary precision integers.
The issue with long double is that it requires 80 bits, and this is
not a multiple of its alignment. This gives a primitive type for
which getTypeSize differed from getABITypeSize. For arbitrary precision
integers it is even worse: there is the minimum number of bits needed to
hold the type (eg: 36 for an i36), the maximum number of bits that will
be overwriten when storing the type (40 bits for i36) and the ABI size
(i.e. the storage size rounded up to a multiple of the alignment; 64 bits
for i36).
This patch removes getTypeSize (not really - it is still there but
deprecated to allow for a gradual transition). Instead there is:
(1) getTypeSizeInBits - a number of bits that suffices to hold all
values of the type. For a primitive type, this is the minimum number
of bits. For an i36 this is 36 bits. For x86 long double it is 80.
This corresponds to gcc's TYPE_PRECISION.
(2) getTypeStoreSizeInBits - the maximum number of bits that is
written when storing the type (or read when reading it). For an
i36 this is 40 bits, for an x86 long double it is 80 bits. This
is the size alias analysis is interested in (getTypeStoreSize
returns the number of bytes). There doesn't seem to be anything
corresponding to this in gcc.
(3) getABITypeSizeInBits - this is getTypeStoreSizeInBits rounded
up to a multiple of the alignment. For an i36 this is 64, for an
x86 long double this is 96 or 128 depending on the OS. This is the
spacing between consecutive elements when you form an array out of
this type (getABITypeSize returns the number of bytes). This is
TYPE_SIZE in gcc.
Since successive elements in a SequentialType (arrays, pointers
and vectors) need to be aligned, the spacing between them will be
given by getABITypeSize. This means that the size of an array
is the length times the getABITypeSize. It also means that GEP
computations need to use getABITypeSize when computing offsets.
Furthermore, if an alloca allocates several elements at once then
these too need to be aligned, so the size of the alloca has to be
the number of elements multiplied by getABITypeSize. Logically
speaking this doesn't have to be the case when allocating just
one element, but it is simpler to also use getABITypeSize in this
case. So alloca's and mallocs should use getABITypeSize. Finally,
since gcc's only notion of size is that given by getABITypeSize, if
you want to output assembler etc the same as gcc then getABITypeSize
is the size you want.
Since a store will overwrite no more than getTypeStoreSize bytes,
and a read will read no more than that many bytes, this is the
notion of size appropriate for alias analysis calculations.
In this patch I have corrected all type size uses except some of
those in ScalarReplAggregates, lib/Codegen, lib/Target (the hard
cases). I will get around to auditing these too at some point,
but I could do with some help.
Finally, I made one change which I think wise but others might
consider pointless and suboptimal: in an unpacked struct the
amount of space allocated for a field is now given by the ABI
size rather than getTypeStoreSize. I did this because every
other place that reserves memory for a type (eg: alloca) now
uses getABITypeSize, and I didn't want to make an exception
for unpacked structs, i.e. I did it to make things more uniform.
This only effects structs containing long doubles and arbitrary
precision integers. If someone wants to pack these types more
tightly they can always use a packed struct.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43620 91177308-0d34-0410-b5e6-96231b3b80d8
in CodeExtractor and LoopSimplify unnecessary.
Hartmut, could you confirm that this fixes the issues you were seeing?
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43115 91177308-0d34-0410-b5e6-96231b3b80d8
Add a new DenseMapInfo::isEqual method to allow clients to redefine
the equality predicate used when probing the hash table.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42042 91177308-0d34-0410-b5e6-96231b3b80d8
In the old way, we computed and inserted phi nodes for the whole IDF of
the definitions of the alloca, then computed which ones were dead and
removed them.
In the new method, we first compute the region where the value is live,
and use that information to only insert phi nodes that are live. This
eliminates the need to compute liveness later, and stops the algorithm
from inserting a bunch of phis which it then later removes.
This speeds up the testcase in PR1432 from 2.00s to 0.15s (14x) in a
release build and 6.84s->0.50s (14x) in a debug build.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40825 91177308-0d34-0410-b5e6-96231b3b80d8
to the worklist, and handling the last one with a 'tail call'. This speeds
up PR1432 from 2.0578s to 2.0012s (2.8%)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40822 91177308-0d34-0410-b5e6-96231b3b80d8
faster than with the 'local to a block' fastpath. This speeds
up PR1432 from 2.1232 to 2.0686s (2.6%)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40818 91177308-0d34-0410-b5e6-96231b3b80d8
to increment NumLocalPromoted, and didn't actually delete the
dead alloca, leading to an extra iteration of mem2reg.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40817 91177308-0d34-0410-b5e6-96231b3b80d8
stored value was a non-instruction value. Doh.
This increase the # single store allocas from 8982 to 9026, and
speeds up mem2reg on the testcase in PR1432 from 2.17 to 2.13s.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40813 91177308-0d34-0410-b5e6-96231b3b80d8
1. Check for revisiting a block before checking domination, which is faster.
2. If the stored value isn't an instruction, we don't have to check for domination.
3. If we have a value used in the same block more than once, make sure to remove the
block from the UsingBlocks vector. Not doing so forces us to go through the slow
path for the alloca.
The combination of these improvements increases the number of allocas on the fastpath
from 8935 to 8982 on PR1432. This speeds it up from 2.90s to 2.20s (31%)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40811 91177308-0d34-0410-b5e6-96231b3b80d8
a using block from the list if we handle it. Not doing this caused us
to not be able to promote (with the fast path) allocas which have uses (whoops).
This increases the # allocas hitting this fastpath from 4042 to 8935 on the
testcase in PR1432, speeding up mem2reg by 2.6x
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40809 91177308-0d34-0410-b5e6-96231b3b80d8
to Instruction::mayWriteToMemory, fixing a FIXME, and helping
various places that call mayWriteToMemory directly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40533 91177308-0d34-0410-b5e6-96231b3b80d8
This interface allows clients to inline bunch of functions with module
level call graph information.:wq
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40486 91177308-0d34-0410-b5e6-96231b3b80d8
second part dominates all the blocks dominated
by original basic block. And first part dominates
second part.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40035 91177308-0d34-0410-b5e6-96231b3b80d8
llvm-gcc build to succeed. Without this change it fails in libstdc++
compilation. This causes no regressions in dejagnu tests. However,
someone who knows this code better might want to review it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@39924 91177308-0d34-0410-b5e6-96231b3b80d8
