This patch adds a feature flag to avoid unaligned 32-byte load/store AVX codegen
for Sandy Bridge and Ivy Bridge. There is no functionality change intended for
those chips. Previously, the absence of AVX2 was being used as a proxy to detect
this feature. But that hindered codegen for AVX-enabled AMD chips such as btver2
that do not have the 32-byte unaligned access slowdown.
Performance measurements are included in PR21541 ( http://llvm.org/bugs/show_bug.cgi?id=21541 ).
Differential Revision: http://reviews.llvm.org/D6355
llvm-svn: 222544
shuffle lowering to allow much better blend matching.
Specifically, with the new structure the code seems clearer to me and we
correctly can hit the cases where merging two 128-bit lanes is a clear
win and can be shuffled cheaply afterward.
llvm-svn: 222539
offsets for code models other than small/medium. For JIT application,
memory layout is less controlled and can result in truncations
otherwise.
Patch from Akos Kiss.
Differential Revision: http://reviews.llvm.org/D6079
llvm-svn: 222538
a bunch more improvements.
Non-lane-crossing is fine, the key is that lane merging only makes sense
for single-input shuffles. Not sure why I got so turned around here. The
code all works, I was just using the wrong model for it.
This only updates v4 and v8 lowering. The v16 and v32 lowering requires
restructuring the entire check sequence.
llvm-svn: 222537
Before this patch, the DAGCombiner only tried to convert build_vector dag nodes
into shuffles if all operands were either extract_vector_elt or undef.
This patch improves that logic and teaches the DAGCombiner how to deal with
build_vector dag nodes where one or more operands are zero. A build_vector
dag node with some zero operands is turned into a shuffle only if the resulting
shuffle mask is legal for the target.
llvm-svn: 222536
lanes.
By special casing these we can often either reduce the total number of
shuffles significantly or reduce the number of (high latency on Haswell)
AVX2 shuffles that potentially cross 128-bit lanes. Even when these
don't actually cross lanes, they have much higher latency to support
that. Doing two of them and a blend is worse than doing a single insert
across the 128-bit lanes to blend and then doing a single interleaved
shuffle.
While this seems like a narrow case, it kept cropping up on me and the
difference is *huge* as you can see in many of the test cases. I first
hit this trying to perfectly fix the interleaving shuffle patterns used
by Halide for AVX2.
llvm-svn: 222533
This patch simplifies the logic that combines a pair of shuffle nodes into
a single shuffle if there is a legal mask. Also added comments to better
describe the algorithm. No functional change intended.
llvm-svn: 222522
E.g., ( a / D; b / D ) -> ( recip = 1.0 / D; a * recip; b * recip)
A hook is added to allow the target to control whether it needs to do such combine.
Reviewed in http://reviews.llvm.org/D6334
llvm-svn: 222510
This mirrors r222331, which enabled SeparateConstOffsetFromGEP on AArch64, in
the PowerPC backend. Yields, on a POWER7 machine, a 30% speedup on
SingleSource/Benchmarks/Shootout/nestedloop (this might just be from LICM,
there is a store moved out of the inner loop) and a potential speedup on
MultiSource/Benchmarks/mediabench/mpeg2/mpeg2dec/mpeg2decode. Regardless, it
makes some code look cleaner, and synchronizing the backends in this regard
seems like a generally good thing.
llvm-svn: 222504
The alloca's type is irrelevant, only those types which are used in a
load or store of the exact size of the slice should be considered.
This manifested as an assertion failure when we compared the various
types: we had a size mismatch.
This fixes PR21480.
llvm-svn: 222499
These recently all grew a unique_ptr<TargetLoweringObjectFile> member in
r221878. When anyone calls a virtual method of a class, clang-cl
requires all virtual methods to be semantically valid. This includes the
implicit virtual destructor, which triggers instantiation of the
unique_ptr destructor, which fails because the type being deleted is
incomplete.
This is just part of the ongoing saga of PR20337, which is affecting
Blink as well. Because the MSVC ABI doesn't have key functions, we end
up referencing the vtable and implicit destructor on any virtual call
through a class. We don't actually end up emitting the dtor, so it'd be
good if we could avoid this unneeded type completion work.
llvm-svn: 222480
Code seems cleaner and easier to understand this way
This is basically r222416, after fixes for MSVC lack of standard
support, and a few cleaning (got rid of a warning).
Thanks Nakamura Takumi and Nico Weber for the MSVC fixes.
llvm-svn: 222472
Currently LoopUnroll generates a prologue loop before the main loop
body to execute first N%UnrollFactor iterations. Also, this loop is
used if trip-count can overflow - it's determined by a runtime check.
However, we've been mistakenly optimizing this loop to a linear code for
UnrollFactor = 2, not taking into account that it also serves as a safe
version of the loop if its trip-count overflows.
llvm-svn: 222451
Windows itanium targets the MSVCRT, and the stack probe symbol is provided by
MSVCRT. This corrects the emission of stack probes on i686-windows-itanium.
llvm-svn: 222439
As dump() methods should be. To allow that, do not store the DWARFFormValue
objects used for the dump in the header data.
Per Alexey's suggestion!
llvm-svn: 222436
These fields would need to be explicitly deleted before we RAUW the temporary
node anyway (this was done in cfe commit r222373). Instead, do not create
these useless nodes in the first place.
llvm-svn: 222434
"global-init", "global-init-src" and "global-init-type" were originally
used to blacklist entities in ASan init-order checker. However, they
were never documented, and later were replaced by "=init" category.
Old blacklist entries should be converted as follows:
* global-init:foo -> global:foo=init
* global-init-src:bar -> src:bar=init
* global-init-type:baz -> type:baz=init
llvm-svn: 222401