llvm-svn: 52226

For now, avoid generating FP select instructions in order to speculatively execute integer arithmetic instructions. FP selects are more likely to be expensive (even compared to branch on fcmp). This is not a wonderful solution but I rather err on the side of conservative.

This fixes the heapsort performance regressions.

llvm-svn: 52224

Avoid duplicating loop header which leads to unnatural loops (and just seem like general badness to me, likely to cause code explosion).
Patch by Florian Brandner.

llvm-svn: 52223

useless insert-extract chains, similar to how it folds them for vectors.

Add a testcase for this.

llvm-svn: 52217

llvm-svn: 52212

llvm-svn: 52191

llvm-svn: 52138

result of a weak function.

llvm-svn: 52137

of apint codegen failure is the DAG combiner doing
the wrong thing because it was comparing MVT's using
< rather than comparing the number of bits.  Removing
the < method makes this mistake impossible to commit.
Instead, add helper methods for comparing bits and use
them.

llvm-svn: 52098

don't make i1 phis when it won't be possible to eliminate them.

llvm-svn: 52097

Speculatively execute a block when the the block is the then part of a triangle shape and it contains a single, side effect free, cheap instruction. The branch is eliminated by adding a select instruction. i.e.

Turn                                                                                                                                                                                                       
BB:                                                                                                                                                                                                        
    %t1 = icmp                                                                                                                                                                                             
    br i1 %t1, label %BB1, label %BB2                                                                                                                                                                      
BB1:                                                                                                                                                                                                       
    %t3 = add %t2, c                                                                                                                                                                                       
    br label BB2                                                                                                                                                                                           
BB2:                                                                                                                                                                                                       
=>                                                                                                                                                                                                         
BB:                                                                                                                                                                                                        
    %t1 = icmp                                                                                                                                                                                             
    %t4 = add %t2, c                                                                                                                                                                                       
    %t3 = select i1 %t1, %t2, %t3

llvm-svn: 52073

llvm-svn: 52053

and better control the abstraction.  Rename the type
to MVT.  To update out-of-tree patches, the main
thing to do is to rename MVT::ValueType to MVT, and
rewrite expressions like MVT::getSizeInBits(VT) in
the form VT.getSizeInBits().  Use VT.getSimpleVT()
to extract a MVT::SimpleValueType for use in switch
statements (you will get an assert failure if VT is
an extended value type - these shouldn't exist after
type legalization).
This results in a small speedup of codegen and no
new testsuite failures (x86-64 linux).

llvm-svn: 52044

otherwise, opt might crash.

llvm-svn: 52041

crash the opt. Just fix this.

Test case in llvm/test/Transforms/InstCombine/2008-06-05-ashr-crash.ll

llvm-svn: 52003

work and how to replace them into individual values. Also, when trying to
replace an aggregrate that is used by load or store with a single (large)
integer, don't crash (but don't replace the aggregrate either).

Also adds a testcase for both structs and arrays.

llvm-svn: 51997

not other instructions. This fixes a crash with the added testcase.

llvm-svn: 51992

passing its pointer. Fixes test with added testcase.

llvm-svn: 51991

function is passed as an argument instead of called. Also do this check a bit
earlier.

llvm-svn: 51990

deprecated by the GVN and GVNPRE passes.

llvm-svn: 51983

llvm-svn: 51955

llvm-svn: 51947

are the same as in unpacked structs, only field
positions differ.  This only matters for structs
containing x86 long double or an apint; it may
cause backwards compatibility problems if someone
has bitcode containing a packed struct with a
field of one of those types.
The issue is that only 10 bytes are needed to
hold an x86 long double: the store size is 10
bytes, but the ABI size is 12 or 16 bytes (linux/
darwin) which comes from rounding the store size
up by the alignment.  Because it seemed silly not
to pack an x86 long double into 10 bytes in a
packed struct, this is what was done.  I now
think this was a mistake.  Reserving the ABI size
for an x86 long double field even in a packed
struct makes things more uniform: the ABI size is
now always used when reserving space for a type.
This means that developers are less likely to
make mistakes.  It also makes life easier for the
CBE which otherwise could not represent all LLVM
packed structs (PR2402).
Front-end people might need to adjust the way
they create LLVM structs - see following change
to llvm-gcc.

llvm-svn: 51928

Don't crash when we encounter one of these.

llvm-svn: 51915

llvm-svn: 51890

llvm-svn: 51887

llvm-svn: 51864

out of instcombine into a new file in libanalysis.  This also teaches
ComputeNumSignBits about the number of sign bits in a constantint.

llvm-svn: 51863

char a[200];
init(a, a+200);

OR

int a[200];
char* b = (char*)a;
char* c = (char*)a;
foo(b, c);

llvm-svn: 51850

llvm-svn: 51848

the conditions for performing the transform when only the
function declaration is available: no longer allow turning
i32 into i64 for example.  Only allow changing between
pointer types, and between pointer types and integers of
the same size.  For return values ptr -> intptr was already
allowed; I added ptr -> ptr and intptr -> ptr while there.
As shown by a recent objc testcase, changing the way
parameters/return values are passed can be fatal when calling
code written in assembler that directly manipulates call
arguments and return values unless the transform has no
impact on the way they are passed at the codegen level.
While it is possible to imagine an ABI that treats integers
of pointer size differently to pointers, I don't think LLVM
supports any so the transform should now be safe while still
being useful.

llvm-svn: 51834

llvm-svn: 51819

llvm-svn: 51817

llvm-svn: 51816

llvm-svn: 51789

Since LCSSA switched over to DenseMap, we have to be more careful to avoid iterator invalidation.  Fixes PR2385.

llvm-svn: 51777

llvm-svn: 51770

llvm-svn: 51698

llvm-svn: 51680

the one case that ADCE catches that normal DCE doesn't: non-induction variable
loop computations.

This implementation handles this problem without using postdominators.

llvm-svn: 51668