llvm-svn: 193957

Instead of doing a RPO traversal of the whole function remember the blocks
containing gathers (typically <= 2) and scan them in dominator-first order.

The actual CSE is still quadratic, but I'm not confident that adding a
scoped hash table here is worth it as we're only looking at the generated
instructions and not arbitrary code.

llvm-svn: 193956

This reverts commit r193356, it caused PR17781.

A reduced test case covering this regression has been added to the test suite.

llvm-svn: 193955

llvm-svn: 193954

llvm-svn: 193947

This adds an SimplifyLibCalls case which converts the special __sinpi and
__cospi (float & double variants) into a __sincospi_stret where appropriate to
remove duplicated work.

Patch by Tim Northover

llvm-svn: 193943

rdar://12856873
Patch by Evan Cheng, with a fix for rdar://13209539 by Tilmann Scheller

llvm-svn: 193942

llvm-svn: 193941

When LLVM is embedded in a larger application, it's not OK for LLVM to intercept crashes.  LLVM already has 
the ability to disable this functionality.  This patch exposes it via the C API.

llvm-svn: 193937

llvm-svn: 193933

llvm-svn: 193932

This is a fixed version of 193928 which keeps these uses in sync.

llvm-svn: 193931

Investigating a bot failure.

This reverts commit r193928.

llvm-svn: 193929

llvm-svn: 193928

llvm-svn: 193927

Doing this with a hash map doesn't change behavior and avoids calling
isIdenticalTo O(n^2) times. This should probably eventually move into a utility
class shared with EarlyCSE and the limited CSE in the SLPVectorizer.

llvm-svn: 193926

There is still a long way to go for llvm-nm, but at least we now match
nm's letter output in the cases we test for.

llvm-svn: 193912

- When selecting BLEND from vselect, the operands need swapping as due to the
  difference between vselect and SSE/AVX's BLEND insn

llvm-svn: 193900

Added -o option to llvm-cov. If no output file is specified, it defaults
to STDOUT.

llvm-svn: 193899

llvm-svn: 193895

llvm-svn: 193894

When the loop vectorizer was part of the SCC inliner pass manager gvn would
run after the loop vectorizer followed by instcombine. This way redundancy
(multiple uses) were removed and instcombine could perform scalarization on the
induction variables. Having moved the loop vectorizer to later we no longer run
any form of redundancy elimination before we perform instcombine. This caused
vectorized induction variables to survive that did not before.

On a recent iMac this helps linpack back from 6000Mflops to 7000Mflops.

This should also help lpbench and paq8p.

I ran a Release (without Asserts) build over the test-suite and did not see any
negative impact on compile time.

radar://15339680

llvm-svn: 193891

In a failed attempt to allow the gnu-public-names.ll test case to not
hardcode the size of the unit that the pubnames section referred to I've
at least managed to have unit headers and pubnames headers print out in
a similar style.

This failed to achieve the desired goal because the header in a unit
specifies the length of the unit without the length element of the
header whereas the length in the pubnames includes this element, so the
numbers are off by 4 bytes. I don't know of any arithmetic powers in
FileCheck so the test case can't simply say "CU_LENGTH + 4".

llvm-svn: 193872

llvm-svn: 193871

llvm-svn: 193869

llvm-svn: 193867

llvm-svn: 193866

linkonce_odr_auto_hide was in incomplete attempt to implement a way
for the linker to hide symbols that are known to be available in every
TU and whose addresses are not relevant for a particular DSO.

It was redundant in that it all its uses are equivalent to
linkonce_odr+unnamed_addr. Unlike those, it has never been connected
to clang or llvm's optimizers, so it was effectively dead.

Given that nothing produces it, this patch just nukes it
(other than the llvm-c enum value).

llvm-svn: 193865

Commenting out this assert because it is causing the build bots to fail.  This effectively reverts r193861, but needs to be fixed as part of r193769.

llvm-svn: 193862

llvm-svn: 193861

If we have a pointer to a single-element struct we can still build wide loads
and stores to it (if there is no padding).

llvm-svn: 193860

Add a Virtualization ARM subtarget feature along with adding proper build
attribute emission for Tag_Virtualization_use (encodes Virtualization and
TrustZone) and Tag_MPextension_use.

Also rework test/CodeGen/ARM/2010-10-19-mc-elf-objheader.ll testcase to
something that is more maintainable. This changes the focus of this
testcase away from testing CPU defaults (which is tested elsewhere), onto
specifically testing that attributes are encoded correctly.

llvm-svn: 193859

Fix Tag_ABI_HardFP_use build attribute to handle single precision FP,
replace deprecated Tag_ABI_HardFP_use value of 3 with 0 and also add
some tests for Tag_ABI_VFP_args.

llvm-svn: 193856

This adds another heuristic to BPI, similar to the existing heuristic that
considers (x == 0) unlikely to be true. As suggested in the PACT'98 paper by
Deitrich, Cheng, and Hwu, -1 is often used to indicate an invalid index, and
equality comparisons with -1 are also unlikely to succeed. Local
experimentation supports this hypothesis: This yields a 1-2% speedup in the
test-suite sqlite benchmark on the PPC A2 core, with no significant
regressions.

llvm-svn: 193855

When a dependence check fails we can still try to vectorize loops with runtime
array bounds checks.

This helps linpack to vectorize a loop in dgefa. And we are back to 2x of the
scalar performance on a corei7-avx.

radar://15339680

llvm-svn: 193853

Clear all data structures when resetting the RuntimeCheck data structure.

No test case. This was exposed by an upcomming change.

llvm-svn: 193852

Drive by space optimization. Also makes the DIEs more regular which
might speed up DWARF parsing.

llvm-svn: 193835

Objective-C data structures.

This is allows tools such as darwin's otool(1) that uses the
LLVM disassembler take a pointer value being loaded by
an instruction and add a comment to what it is being referenced
to make following disassembly of Objective-C programs
more readable.

For example disassembling the Mac OS X TextEdit app one
will see comments like the following:

movq    0x20684(%rip), %rsi ## Objc selector ref: standardUserDefaults
movq    0x21985(%rip), %rdi ## Objc class ref: _OBJC_CLASS_$_NSUserDefaults
movq    0x1d156(%rip), %r14 ## Objc message: +[NSUserDefaults standardUserDefaults]
leaq    0x23615(%rip), %rdx ## Objc cfstring ref: @"SelectLinePanel"
callq   0x10001386c ## Objc message: -[[%rdi super] initWithWindowNibName:]

These diffs also include putting quotes around C strings
in literal pools and uses "symbol address" in the comment
when adding a symbol name to the comment to tell these
types of references apart:

leaq	0x4f(%rip), %rax ## literal pool for: "Hello world"
movq    0x1c3ea(%rip), %rax ## literal pool symbol address: ___stack_chk_guard

Of course the easy changes are in the LLVM disassembler and
the hard work is up to the implementer of the SymbolLookUp()
call back.

rdar://10602439

llvm-svn: 193833

llvm-svn: 193823

llvm-svn: 193819