In AVX 256bit vectors are valid vectors and therefore the Type Legalizer doesn't
split the VSELECT and SETCC nodes. AVX only supports MIN/MAX on 128bit vectors
and this fix enables vector splitting for this special case in the X86 DAG
Combiner.
This fix is related to PR16695, PR17002, and <rdar://problem/14594431>.
llvm-svn: 191131
The Type Legalizer recognizes that VSELECT needs to be split, because the type
is to wide for the given target. The same does not always apply to SETCC,
because less space is required to encode the result of a comparison. As a result
VSELECT is split and SETCC is unrolled into scalar comparisons.
This commit fixes the issue by checking for VSELECT-SETCC patterns in the DAG
Combiner. If a matching pattern is found, then the result mask of SETCC is
promoted to the expected vector mask for the given target. This mask has usually
te same size as the VSELECT return type (except for Intel KNL). Now the type
legalizer will split both VSELECT and SETCC.
This allows the following X86 DAG Combine code to sucessfully detect the MIN/MAX
pattern. This fixes PR16695, PR17002, and <rdar://problem/14594431>.
llvm-svn: 191130
A patch to AllocateTarget function to recognize llvm::Triple::NaCl for
MIPSEL and return NaClTargetInfo. Additional test has been added to check
if the expected macros get defined.
llvm-svn: 191124
Reapply r191108 with a fix for a memory corruption error I introduced. Of
course, we can't reference the scalars that we replace by vectorizing and then
call their eraseFromParent method. I only 'needed' the scalars to get the
DebugLoc. Just store the DebugLoc before actually vectorizing instead. As a nice
side effect, this also simplifies the interface between BoUpSLP and the
HorizontalReduction class to returning a value pointer (the vectorized tree
root).
radar://14607682
llvm-svn: 191123
The problem of r191017 is that when GVN fabricate a val-number for a dead instruction (in order
to make following expr-PRE happy), it forget to fabricate a leader-table entry for it as well.
llvm-svn: 191118
Clean up some simple code quality issues. Bring internal naming
conventions up to current standard, fix inconsistent formatting, and
tidy up a couple of odd contructs.
llvm-svn: 191117
Match reductions starting at binary operation feeding into a phi. The code
handles trees like
r += v1 + v2 + v3 ...
and
r += v1
r += v2
...
and
r *= v1 + v2 + ...
We currently only handle associative operations (add, fadd fast).
The code can now also handle reductions feeding into stores.
a[i] = v1 + v2 + v3 + ...
The code is currently disabled behind the flag "-slp-vectorize-hor". The cost
model for most architectures is not there yet.
I found one opportunity of a horizontal reduction feeding a phi in TSVC
(LoopRerolling-flt) and there are several opportunities where reductions feed
into stores.
radar://14607682
llvm-svn: 191108
is no need to go through the driver indirection here, and it clutters
things up as dependencies can sneak in for specific things the driver is
doing.
llvm-svn: 191107
Specifically, allows the unwinder to handle the case where sc.function
gets resolved with a pc that is one past the address range of the function
(consistent with a tail call). However, there is no matching symbol.
Adds eSymbolContextTailCall to provide callers with control over the scope
of symbol resolution and to allow ResolveSymbolContextForAddress to handle
tail calls since this routine is common to unwind and disassembly.
llvm-svn: 191102
This solves the problem of fallback onto ourselves if clang-cl
has been renamed to cl.exe and put on the PATH, as happens with
the VS integration.
Differential Revision: http://llvm-reviews.chandlerc.com/D1731
llvm-svn: 191099
Apart from being more compact and already implemented, this also handles the
case where the parent is null. (It does also ignore all casts, not just
implicit ones, but this is more efficient to test and in the case we care
about---a message in a PseudoObjectExpr---there should only be implicit casts
anyway.
This should fix our internal buildbot.
llvm-svn: 191094
The GEP pattern is what SCEV expander emits for "ugly geps". The latter is what
you get for pointer subtraction in C code. The rest of instcombine already
knows how to deal with that so just canonicalize on that.
llvm-svn: 191090